Item — HN Reader

> Once launch costs are down to $50/kg or so, whatever loss we have in earthbound astronomy can be more than compensated by large telescopes in orbit and on the far side of the moon.

I don't think launch costs are the dominant expense in space telescopes. I think most of the expense is technology development for one-off (or small run) instruments as well as paying the people who design and build it, write the software, and operate the telescope. So even if the launch were free, there's still a huge expense in building space-based telescopes. I believe the rule of thumb is space-based costs ~10x that of a ground based telescope.

The largest astronomical mirror in space is 3.5m (Herschel). 30m optical telescopes are being built on the ground now and 10m telescopes have existed since the early 1990s. It would be incredibly expensive to put a 30m telescope in space, even if the launch were free.

3 comments

The price for space telescopes will come down dramatically once you no longer have to arrange for them to assemble themselves in orbit, or wait for surplus hills from spy agencies.

You can lose all the systems used on Earth to compensate for atmospheric and gravitational effects, and launch it to an orbit well above any satellite constellation to eliminate satellite trails from the picture.

With the advent of heavy and super heavy launchers there will no longer be an absolute focus on reducing weight. This alone will bring the cost of space telescopes by an order of magnitude.

Space telescopes also completely avoid the problem of facility construction impinging on indigenous sacred sites.

Unless you view the sky as an (indigenous) sacred site.

Very few satellites impact the naked eye night sky.

A space telescope would by its very nature be less impactful than most satellites, since it will be placed in a high orbit away from the glare of the Earth, also ensuring that the Earth fills less of the sky to allow better access for astronomy.

Space hardware costs 10X as much as ground partly because launch is very expensive. When launch is so expensive, it's worthwhile building your satellite to be extremely lightweight, but also extremely robust, since it costs so much to launch a replacement.

Starship has a 9-meter cargo space. Musk recently proposed a successor with twice that diameter. What could we do with a space telescope that size?

The real win though is lunar telescopes. Then it is ground-based. You can build it from raw materials on the moon, which is 45% silica and also has plenty of iron and aluminum. You've got no atmosphere and 1/6 gravity; you can make it huge and get perfect clarity.

This may sound far-fetched but the whole point of SpaceX is to set up permanent human habitation on the moon and Mars, in the fairly near future.

There might not be atmosphere, but lunar dust is notoriously abrasive and also "sticky" because of electrostatic charge.

Those are serious issues that would have to be resolved before an investment in a telescope should be made.

Well the obvious question is how much cheaper would space telescopes be if they mass produced and launched a standard design instead of one-offs?

What if they took the JWST design ($10 billion in development costs!) and made 10 or 100 more copies? How much would each additional copy cost to build and launch?

> What if they took the JWST design ($10 billion in development costs!) and made 10 or 100 more copies? How much would each additional copy cost to build and launch?

There's presumably a way to estimate that. But you'd also want to ask if it's even useful to have 10 or 100 copies of JWST? Even if the unit cost is lower, the total program cost would still be greater than the $10b for one. The astronomical community is probably not large enough to fully utilize 10 JWSTs, let alone 100. I'm sure we could come up with enough observations to keep them all busy, but there isn't enough person-power to analyze and interpret all that data. So if one has money to build 10 JWST's, a likely better use of that money would be to build one or two and spend the rest of the money hiring people (graduate students, postdocs, faculty) to do science with the data. Or just build one JWST and use the money for a second to build something else that pushes capability in other areas (e.g., the "Origins Space Telescope" far-infrared mission concept, or the Lynx X-ray telescope concept). I suspect that would be better for how much new science you can do per dollar spent.

I think the answer to that is “not very much.” Telescopes use a lot of highly specialized parts made of rare and expensive materials. They use a lot of gold for various components and the mirrors are made of gold or platinum. The tolerances in the fabrication of the mirrors are EXTREMELY tight in order to reduce aberrations as much as possible. This sort of high precision manufacturing with exotic materials simply doesn’t scale.

"EXTREMELY tight"

Telescope mirrors need to be polished to an accuracy of a quarter-wavelength of whatever radiation you want to capture - so for an optical telescope, that's a quarter-wavelength of visible light. Any greater precision than that is spurious, and won't get you better imaging. Less precision than that == bad mirror.

The techniques for polishing and testing small mirrors (up to say 12") are amusingly low-tech, almost steampunk. You can get to 1/4 wavelength of visible light with a homemade grinding/polishing rig, and a homemade optical bench that consists of little more than a jig to hold the mirror, a lamp and a couple of razor blades.

I'm sure the challenges in building accurate 30m optical mirrors are massively bigger, not least because very big mirrors warp under their own weight. But I imagine the principles are just the same.

Incidentally, I had no idea people were building 30m ground-based optical telescopes; that's just insane. I mean, my house is only 8m high.