Hacker News Reader

Comment by robcohen | Hacker News Reader

I understand that the astronomers are disappointed, but isn’t that because the period we are in right now is immediately before it becomes 10x cheaper to launch telescopes into space? Won’t it be so much better to observe directly from space as opposed to from the earth?

Can anyone comment on this?

View on HN

Not since the advent of modern adaptive optics[1]. Ground-based telescopes now outperform space telescopes in the visible spectrum[2]. For as long as the cost per kilo to orbit is higher than the cost per kilo to the Atacama desert, ground-based astronomy is going to be superior.

The other great advantage of ground-based telescopes is that _they're easy to fix and improve_. The iteration loop is that you walk into the dome and you fiddle with things. You don't need multiple years of planning to polish a mirror. There's an excellent Twitter thread on this issue here[3].

[1] https://en.wikipedia.org/wiki/Adaptive_optics

[2] https://astronomynow.com/2018/07/19/adaptive-optics-tomograp...

[3] https://twitter.com/cynosurae/status/1196567663977230336

Perhaps a deal could be worked out where the satellite companies also have to launch Atacama-scale telescopes, where Musk et al have to pay for the difference in cost between Atacama and orbit.

Telescopes are pretty big. It would take 5-10 Falcon 9 launches to put something with that mass into orbit (although in space you could probably have a lighter mechanism for orienting el/az). The telescope's mirror is also too big to physically fit into the F9's fairing.

Once you get it there, you need to assemble it and keep it running. Maintenance is a real doozy— building a telescope on Earth is expensive. Building a telescope in space is even harder, since it won't be regularly serviced and you have to overengineer it. A typical space telescope still costs an order of magnitude more than a recordbreakingly expensive earth-based one.

Yes but the point of Starlink is to fund Starship, which is a lot bigger than Falcon 9. Its fairing is 9 meters in diameter[0], large enough for all but the top six land-based telescopes.[1] Its payload is 100+ tonnes, five times larger than Falcon 9's.

Starship is also a lot cheaper per kilogram of payload; at scale, as low as $20/kg. Cheap launch also makes your satellites cheaper to build, since it's less of a disaster if you have to replace them.

Musk has already floated the idea of building a second-generation Starship with an 18-meter fairing.[2]

[0] https://www.spacex.com/starship

[1] https://en.wikipedia.org/wiki/List_of_largest_optical_reflec...

[2] https://www.teslarati.com/spacex-elon-musk-starship-the-next...

Even if it becomes physically possible to launch them at some point in the future, I should have mentioned that launch costs are only a relatively small part of the cost of space based telescopes. As launch prices continue to go down, this will be more and more true in the future. For the upcoming JWST, the launch is only costing around 1-2% of the overall price. It will be a similar number for WFIRST if that gets built on budget. The expensive part is engineering a device that can assemble itself in outer space, and continue to run 24/7 for decades without regular maintainence.

You don't need much orbital assembly if you can launch something with a 9-meter or 18-meter diameter. And again, cheaper launch cost also makes space hardware cheaper; you can afford to make it heavier, and you can afford to replace it if necessary.

This is outdated model regarding space telescopes.

With radically cheaper launches, it's becoming quite possible - not now, but somewhat soon - to actually service telescopes, with conditions superior to those under which the Hubble was serviced by Shuttle teams.

With radically cheaper launches, it's not necessary to ensure the 'scope runs 24/7 for decades without regular maintenance. It can even be replaced regularly; including by bringing it to Earth, servicing it and sending back.

Launch prices are going to drop, but so is the necessity to be extremely reliable for the space equipment, for the same two reasons: 1) it's comparatively easy to get to it for a service and 2) it's comparatively easy to just send a replacement if needed.

With space manufacturing advancements still ahead of us - we didn't really experimented with making telescopes in orbit - it's not clear if space telescopes won't outperform terran ones. Space telescopes are going to have advantages like "atmosphere doesn't block part of the spectrum" and "there is no need to make a big structure capable withstanding 1g gravity".

It's still crucially important to be able to do terran-based astronomy, and that's a problem, but there is no doubt that the astronomy revolution which started with Hubble still has major advancements to have in the near future.

Also we are gone move into types of systems where its not one thing anymore. We increasingly telescopes that are consistent of multiple parts sometimes including things like a sun shade that operates far away from the rest of the parts.

All of these things are only gone be possible once launch prices are cheaper and Starlink (or something like it) is a major part of how that can happen.

Radically cheaper launches will still be significantly more expensive than current maintenance contracts. Even if we have infinitely reusable rockets in the future, you're still looking at millions of $ in fuel costs for a servicing trip. I am personally not as optimistic about "soon".

If you send a technician out to the Earth-Sun L2 point (where JWST will be) so that she can throw some grease on a gear or spray WD-40 on a reaction wheel, she'll be 4 times further from Earth than any human has ever gone. Although there may be some synergy with future Lunar/Martian missions, there's a lot of work that needs to be done to figure out how to do that safely. I think L2 is the safest Lagrange point w.r.t radiation, but it's still potentially deadly.

>"there is no need to make a big structure capable withstanding 1g gravity".

While the overall structure may not need to withstand 1g continuously, you still need to design it to as much as 6 Gs of acceleration during launch (as well as significant lateral acceleration, and occasional shocks in the dozens or hundreds of gs).

Adaptive optics should be able to adapt to any number of satellites traveling, quite predictably and diminutively through their frame of capture.

Signal to noise ratio is the limiting factor in these observations. Decrease the signal and suddenly whole sets of galaxies fall below the noise floor.

I have not seen any information that light from these satellites could scatter and brighten the night sky a significant amount. This article just shows their paths are visible on a sensitive image, but the paths themselves amount to tiny fraction of the image. There is no claim they added significantly to the overall exposure. With their being bright planets and stars already present in the sky, I don't expect they could.

That is not remotely what "adaptive optics" means. Word arguments like that degrade the quality of debate.

In case the commenter isn't aware, adaptive optics compensate for the shimmer of the atmosphere, not the presence of light-scattering objects between the telescope and a dim object part-way across the universe.

Relevant light scattering occurs in the atmosphere between the satellite and the telescope. It also occurs between every other light source, across the universe, and the telescope. Thats why the satellites contribution to light scattering is tiny. Even in the morning where one might catch the sun - the sky is already lightened.

Hard or soft optical adaptations to remove atmospheric distortions are far more demanding than processing required to remove low magnitude satellite trails, which is already developed and required on occasion.

Im broadly aware of a range of techniques which can be involved in adaptive optics. The quality of debate seems very low to me and not worth laboring words on when people are imagining low power satellite trails represent a big challenge to astronomical image clarity.

How much do you think all the sunk infrastructure costs of existing telescopes cost, and how many years did it take to build them?

The bigger telescopes today have primary reflectors that are vastly bigger than the James Webb Space Telescope: for example the Large Binocular Telescope with its 8.4 metre mirrors, or the Extremely Large Telescope (currently under construction: first light due in 2024) with its 39.2 metre segmented priary. Even the segmented designs weigh hundreds of tons; these instruments are too big and heavy to put in orbit even with SpaceX's Superheavy/Starship.

If the light pollution from Starlink can't be mitigated, it'll be necessary to replace dozens of hundred million dollar to billion dollar instruments which took years or decades to build, at a post potentially greater than the entire consat cluster.

What's more important? Global internet access for the rest of the world population or a few dozen telescopes? Imo internet access is going to have immediate positive impact for billions of global disconnected poor.

I'm sure I could make better use of your home by demolishing it and erecting a homeless hostel. Isn't that more important than your prior claim to live there?

(This is a m-e-t-a-p-h-o-r, not a literal proposal, but what Starlink is doing to astronomy on a global scale is pretty much equivalent. The onus should be on Starlink to (a) not do that, or (b) pay for complete relocation or replacement of the telescopes they've just wrecked. Yes, that's a lot of money. Boo hoo ...)

> The onus should be on Starlink to (a) not do that

Why? Astronomers have no more claim to ownership of the sky than Starlink does.

Astronomers are not changing anything IN the sky, Starlink does?

The sky changes all the time. Astronomers are observers, and it seems the onus should be on the observer to change the way they interpret the observation if the observable changes. We didn't kill radio stations (or microwave ovens) to make sure that radio telescopes can do their jobs [https://xkcd.com/2226/].

Isn't there a "security-zone" around the large radio-telescopes where usage of most electronic devices is forbidden?

> Imo internet access is going to have immediate positive impact for billions of global disconnected poor.

do you really think that? is that why companies like facebook are so interested in third-world markets? or is it because they want to elevate their ability to siphon user data and hack human emotions to become addicted for more and more money, and also that facebook wants there to be little distinction between them and and the internet in these markets?

i don't think the internet has had as big of a positive impact on society as people think it has. in no other time in history has there existed such a powerful medium for control and manipulation.

and this is ignoring the fact that most of elon's projects are born out of narcissism and hubris.

Why not have both?

Space telescopes are objectively superior in every single way. It's like a photograph under clear conditions vs a photograph in fog. Look at Hubble vs much larger terrestrial scopes. Starship's booster could launch something much larger than Hubble at a fraction of the cost.

That's not really true anymore. In the last few years adaptive optics has closed the gap.

source: https://www.popularmechanics.com/space/a22240382/neptune-gro...

And if Starlink inconveniences astronomy, it probably won't take but so many further years before adaptive signal correction algorithms close the gap again.

1) Atmosphere blocks parts of the electromagnetic spectrum, so not all telescopes on Earth will be on the level field with space-based ones.

2) Earth-based telescopes have to withstand gravity of 1g, while in space you can have some bigger structures.

However, sunk costs are sunk costs. There's a reason we have that term in economics in the first place.

It is hard to estimate the cost of replacing LSST with something off-planet. As it, alone, costs ~billions, putting a similar instrument in space costs far, far more.

Furthermore, dark skies are a resource for all of humanity, not just astronomers.

The Starlink project doesn't actually threaten dark skies. The satellites spend most of their night-side transit in Earth's shadow.

Indeed, a lack of dark shy makes the Starlink constellation harder to see from the ground.

That's like saying arson is no problem, the cost of the building was a sunk cost anyway.

Arson adds no value. The value trade-off of slightly decreased telescope efficiency and ubiquitous internet access is a lot more positive.

I can't comment on that specifically (telescopes in orbit, I invite others to). My two cents though are this: I frequently go on multi-day backpacking trips just to get away from it all. I can get so far out when the only human-made anything that I see are the occasional aircraft and the occasional satellite. It's a great disconnect -- I can pretend that the modern world doesn't exist. It would be a great disappointment to me if the evening/morning sky were to become so populated that there were constant reminders of the modern & interconnected world. I would be resentful of whoever ruined that for me, everyone else, and generations to come.

On these trips, do you bring an emergency radio or GPS?

Typically, no, but sometimes yes . It depends on how well I know the area (if at all), how far from the road, how many people are coming with, etc...I have a Yaseu FT-60, but (a) Don't have HAM license and (b) It's too heavy. I never take a GPS map, but sometimes I take an emergency GPS emergency locator (e.g. SPOT).

You won't be able to see these once they are at operational orbit unless you use a telescope. And, as the article says, future launches will be nonreflective.

That is encouraging. Thank you.

On one hand, yes, you're entirely right about the eventual outcome, but on the other hand, let's get there gracefully. The article describes an event akin to commercial light pollution or commercial aviation, with the similar cause of private industry charging ahead of regulation and standards. It could be that the only solution is to quickly act to regulate space more completely, but we don't have a healthy international political climate for such regulation right now.

Also, in terms of quality of evidence, there is sadly a relatively strong pseudoscientific strain of thought that denies any astronomy done from space. If we stop taking high-quality photos from the surface of the Earth using relatively mundane methods, then the flat-Earthers win.

Finally, what if Kessler Syndrome sets in? In that case, it would not be cheaper to launch a telescope into space, even if the telescope is destined for a relatively distant place like a Lagrange point, since it first must run a gauntlet of debris. Meanwhile, observations from the ground would certainly be twinklier and noisier but not especially more expensive.

> On one hand, yes, you're entirely right about the eventual outcome, but on the other hand, let's get there gracefully.

The problem with a "get there gradually" approach (at least as I understand it, and I may not have understood you correctly) is that it pretty much precludes a useful LEO constellation.

A LEO constellation is not really useful unless it has a significant number of satellites. Putting up 1% of the satellites doesn't really get you 1% of the functionality of the system. There's a minimum threshold you have to hit for the system to really be viable at all.

Which means at _some_ point (if we have LEO constellations at all) we're going to have to really rapidly go from very few LEO satellites to a large number.

> in terms of quality of evidence, there is sadly a relatively strong pseudoscientific strain of thought that denies any astronomy done from space. If we stop taking high-quality photos from the surface of the Earth using relatively mundane methods, then the flat-Earthers win.

This isn't a real problem with the quality of the photos. Such people should be kept out of power and otherwise ignored.

> what if Kessler Syndrome sets in?

Starlink can't result in Kessler Syndrome. The time it takes orbits to decay from that altitude is measured in single-digit years.

Of all the political groups that we should be worried about catering to regarding best course of action for deployment of space communication tools, the flat-Earthers should be at the absolute bottom of the concern pile.

Launch costs are not the dominant cost of building space telescopes, which are ~10x more expensive than equivalent ground-based systems (when they exist).

Ground-based telescopes built in the 1940s are still being used for cutting-edge science today because it's simple and cheap to upgrade their instruments. That's not possible in space--when JWST runs out of propellant, it becomes just a $10B piece of space junk.

Moon based telescopes then. Let's do it!

Putting a person on the moon cost a significant fraction of the US GDP.

The tax burden required to move Earth's level of free dark-sky accessibility to space (step outside at night, buy binoculars, a Celestron, or LSST, and just look up) would be crippling for any nation.

But think of the return on the investment of cool.

Observations are being ruined now. Those cheap space telescopes will be available when ecactly? Is there even any project in the works besides JWT? Projects like these take a lot of time.

Improve observation data post processing. Lasers are used as artificial stars [1] to subtract atmospheric distortion, use stacking [2] to remove StarLink tracks. Those observations aren't being "ruined", they simply have some noise to remove.

Long term, we should be lifting observatories (both optical and RF) into orbit, Lagrange points, or the Moon. Observing from the ground sucks comparatively, and there's a lot of space out there.

[1] https://en.wikipedia.org/wiki/Laser_guide_star

[2] https://www.skyandtelescope.com/astronomy-blogs/imaging-foun... (Satellites, Begone! Remove Trails from Your Astrophotography)

That might work for imaging, but what's your fix for spectroscopy?

Optical, RF, or Xray?

Let's start with optical.

Yes WFIRST[0] is currently in the works.

JPLRMA[1] is investigating low cost mirror for space observation although they are primarily focused on the outer planets of our solar system.

[0]https://wfirst.gsfc.nasa.gov/ [1]https://www.nap.edu/resource/13117/App G 19_Enceladus-RMA.pdf

Reply on news.ycombinator.com