Hacker News Reader

Comment by mlyle | Hacker News Reader

Yup. Though--- I do wonder about the desire to keep it functioning at all costs. Servicing is expensive: humans and their consumables aren't cheap.

I wonder what we'd get instead if we dumped equivalent or slightly more resources into a new telescope, instead.

View on HN

We're going to have another Hubble up in 2027 if all goes well: the Nancy Grace Roman Space Telescope, or Roman for short. Originally known as WFIRST, it's a 1.6m-primary-mirror declassified Keyhole spy sat similar to Hubble, but with a shorter focal length and thus a much wider field of view (around 1/4 square degrees, roughly the area of a full moon). Still, its angular resolution will be comparable to Hubble thanks to a much higher-resolution, 300 megapixel imager. It's a visible/near IR instrument, so won't replace Hubble's near UV capabilities. Also, Roman won't have a traditional spectrograph on board. What it will have, however, is a novel coronagraphic instrument for imaging and spectrography of companions of nearby stars (brown dwarfs and gas giants in wide orbits -- still no imaging exo-Earths or even exo-Jupiters, alas).

https://en.wikipedia.org/wiki/Nancy_Grace_Roman_Space_Telesc...

https://roman.gsfc.nasa.gov/why_Roman_Space_Telescope.html

Why do they call it an infrared telescope when it doesn't see in the infrared spectrum?

It does in near IR, but yeah, it’s as much a visible light telescope as an IR one.

It sees the near infrared. I expect they got that for free from the camera sensor. https://en.wikipedia.org/wiki/Infrared_photography#Digital_c...:

“Digital camera sensors are inherently sensitive to infrared light, which would interfere with the normal photography by confusing the autofocus calculations or softening the image (because infrared light is focused differently from visible light), or oversaturating the red channel. Also, some clothing is transparent in the infrared, leading to unintended (at least to the manufacturer) uses of video cameras”

I expect NASA to do the calculations, but I would expect boosting the orbit of an existing working telescope would be much cheaper than building a new one AND ALSO putting it in orbit. There are relatively few telescopes outside Earth's atmosphere; keeping Hubble running seems like a good idea.

It's not just orbital boosting.

It's also doing things like swapping out reaction wheels and dying hardware of various kinds.

And as soon as you're doing that, you're probably using humans. And then because the cost is already so big, you might as well replace some instruments, too...

Do they plan to do more manned missions to Hubble?

Hubble re-enters in 2037, and is useless because of orbital mechanics a few years before that.

I think it's unlikely the current reaction wheel system lasts to that ~2032 timeframe, let alone significantly beyond.

Of 6 wheels, 2 are working well, 1 is dodgy, and 3 have failed. 3 are required to keep orientation.

I thought they could now control it with only 2 reaction wheels [1]

[1] https://core.ac.uk/download/pdf/10543974.pdf

This proposal (by being part of Polaris) is implied to be manned as far as I can tell.

It's also proposed to be funded by Isaacman. Reading between the lines, he's probably planning to go to space personally and fix the Hubble. Which is kinda a vanity project at that point, but hey, I'll take it.

Both he and SpaceX seem committed to expanding the capabilities of private spaceflight and private astronauts. If they are into it, why not?

Yes, humans are needed until they’re not.

Once they’re no longer needed the cost of repair drops by 100x.

Musk’s android seems like a bit of a gimmick but hopefully he stirs up enough imagination in people that we get a few breakthroughs in robotics.

Then we can explore the entire solar system…build bases on Mars and the moons of Jupiter, etc without sending humans

we can't even get a robot to fix a dishwasher or to sort trash. Let alone am orbital telescope

I can't even get my dishwasher to clean my dishes. This is not the future I was promised.

The trick is let your dog lick them clean.

biological automation is always superior

>Where can you find another non-linear servo-mechanism weighing only 150 pounds and having great adaptability, that can be produced so cheaply by completely unskilled labor?

—Scott Crossfield, US test pilot <https://quoteinvestigator.com/tag/albert-scott-crossfield/>

Yeah, many people think it’s impossible until someone else does it.

The game of Go wasn’t close to being won by a computer until DeepMind came along.

https://www.theverge.com/2016/3/9/11185030/google-deepmind-a...

My city removed all plastic recycling bins several years ago in favor of using trash sorting robots after collecting mixed waste.

The robots could pick sort 70% effectively while the population of humans after years of training was stuck at 50%. I imagine the efficiency of the robots may even have increased a bit since then.

Interesting! Which city, and is there a writeup with more details?

You missed crucisl detail- humans were sorting for free, owner of robots is paid.

Another crucial detail: producers of platic packaging engage is deliberate fraud to conflate recycleable and non recycleable plastic. Plastics that say 'commonly recycled' may be unrecycleable anywhere except a leading technology lab.

Another detail: massive fraud in recycling industry:

https://www.theguardian.com/environment/2018/oct/18/uk-recyc...

Most european countries stil do separate recycling, and have much higher recycling rates than UK or US, which have mixed recycling.

To be fair, making a robot to fix a specific kind of dishwasher whose internal layout is known and where fixes involve replacing parts already meant to be replaceable is probably a lot easier than one that can fix any dishwasher. Same with sorting trash.

The big point here is that this boosting and potentially servicing mission isn't being pitched by NASA. It's a SpaceX proposal connected to their entirely private program to develop EVA capability for Dragon and to eventually test Starship's ability to support crews in space. Thus for NASA it would likely be cheaper than if they were asking for it.

So the reason we likely won't see a robotic servicing of Hubble (from this) is that it doesn't have as much relevance to SpaceX's goals with Polaris Dawn as just docking, orbit raising and maybe crewed servicing does.

Agreed, technologically you can make a robot that can fix a dishwasher that would be specially designed to be repairable.

But if you can't get dishwasher manufacturers to produce repairable dishwashers, then it's all for naught - we have failed as am industry.

Amd if you can get manufacturers to produce repairable dishwashers, then probably those dishwashers can be repaired by random joe opening a repairshop.

And its not clear why we need a robot.

Thats why our industry is a joke, we were promised robots will deliver our food, but instead robots micromanage indentured servants.

A boost module can semipermanently attach itself and take over all the maneuvering functions with its own hardware suite.

Ha, found the practical engineer! I didn't think of that. Brilliant!

Or Kerbal Space Program player lmao.

I definitely did that once or twice, just send a small craft with some RCS thrusters and reaction wheels then attach it to spacecraft via The Claw

On the last Hubble service mission, they added a soft capture mechanism. They could attach a boost module to that.

https://en.wikipedia.org/wiki/Hubble_Space_Telescope#/media/...

The original Hubble servicing mission where apparently more expensive than building and launching 3 replacements. Those missions where less about cost than actually testing repair capabilities and some publicity.

Which may or may not still be relevant.

ALL shuttle missions were more expensive than launching 3 separate expendable missions. (on top of the launch costs themselves, the payload had to be 'upgraded' to human safety requirements. So things like deployment squibs had to be made triple redundant (instead of double) and had to be completely disconnected from a power source until released from the shuttle bay. (and of course, that power connection system had to be triple redundant... the complexity fed on itself) and that's before you get to the paperwork and the safety reviews... And you had to staff up three launch teams, instead of two... you needed a full team at Johnson in addition to the Cape and your own site. And the launch teams needed to be onsite for the full shuttle mission and rehearsals, and jerked around by the incessant shuttle schedule slips. The costs just kept piling up. )

As other's have commented, NASA has "free" telescopes donated to them by the NRO.

The next space telescope will be the Nancy Grace Roman telescope which is a wide angle telescope so not an exact replacement.

https://roman.gsfc.nasa.gov/

The longer term goal is the Luvoir telescope but its not yet been approved and will probably require significant R&D like JWST.

https://www.luvoirtelescope.org/

Sure. I'm not talking about the timelines of existing plans for new telescopes. I'm comparing ROI vs:

- mission to reboost and service Hubble in various ways

- launching a new space telescope mission

It takes about fifteen years from concept studies to launch for a Hubble-class space telescope. I guess you could make it in ten if you just chose to replicate Hubble's instruments and if NRO had another spare Keyhole stashed somewhere...

The NRO donated two and parts for a third to NASA in 2012. One is apparently being used for the Roman Space Telescope, which would seem to leave another and a primary mirror of a third still unused.

Perhaps if we keep it in orbit long enough, robot technology will improve to a point where we don't need to send any humans to service the Hubble anymore.

Also, servicing the Hubble is exactly the kind of mission that commercial space ventures can use to showcase their tech. Relatively low risk, high media exposure. A pretty good practice ground, really.

By the time robotics advances to that point, it would most certainly be economical to manufacture a significantly more capable replacement telescope, wouldn't it? As far as I know, there isn't any particularly exotic materials in Hubble, it's just a big heavy precision manufactured (lol) object.

Even right now, it seems like the real cost of a falcon heavy launch is relatively low. We would get more science for the money by mass producing several higher performance telescopes with fewer reliability requirements.

True for now, but Starship, if full and rapid reusability is achieved, will radically change the economics of space missions.

A new hubble-scale space telescope is a multibillion dollar project - Hubble cost about $7 billion in 2022 dollars at the time of launch, JWST cost $10 billion.

By comparison, a manned servicing mission would likely cost in the mid hundreds of millions, or about 1-10% of the cost of a new telescope. It's unlikely that a one time injection of cash would buy much additional functionality for future satellites. By comparison, buying an extra few years before you need a proper replacement satellite would allow you to build a substantially better telescope at the same annual funding rate, or build the same telescope at a lower annual funding rate.

Things only cost that much per unit because they are one-offs. The trick to saving pots of money is to build duplicates, which will not incur research, design, tooling, test rig, etc., costs.

The other problem with these is the big super high quality optical mirrors. Going to a production line with those would be really interesting! I don't think you'd get a big enough economy of scale to mass produce them in a just-in-time pipeline though.

Once you've designed and built the grinding machine (I saw a picture of it, it fills a special room) you've done most of it. Same for the custom test rig.

SpaceX shows what happens to costs when you re-use designs and tooling.

True but there's also decreasing marginal utility for each duplicate. A second telescope in any given class would be awesome but it's always going to be less groundbreaking and useful than the first one.

The universe is so big, and so many things to look at, I find it hard to believe we couldn't productively use 100 Hubbles.

100 Hubble's would be great but 100 Hubble's and nothing else would not be

Well if we had built a warehouse full of Hubbles, then launching a duplicate would make a lot of sense, but we didn't and now the cost of making a replacement after all the custom tooling and talent has been lost is just as expensive as the first time around.

The next gen space telescopes are already being planned. It just takes a long time to get them completed. In the mean time, we have a working telescope. We should just let it die and not be used while we wait for the next to come about?

Older telescopes can still contribute to science in parallel, too. Even terrestrial telescope time is booked years in advance. More scopes means more time for secondary pursuits, hands-on time for more junior astronomers, observations for more speculative theories, etc.

Terrestrial telescopes don't have nearly the degree of maintenance costs that a manned servicing mission in orbit entails.

If fixing your old terrestrial 0.5m telescope were going to cost more than a shiny new PlaneWave 0.5m, you can bet that you're going to replace.

Of course, the cost tradeoff is quite different. All I'm saying is that a telescope, ground or space, does not suddenly become irrelevant because it's not the latest and greatest, so the expected value of the Hubble may remain positive after newer scopes are launched.

Each new space based platform also is built to see different parts of the spectrum. Hubble happens to be mainly visible. It is not irrelevant just because JWST is now active. To the contrary, they are combining images from each platform as a composite because they see different things for the same objects.

There would be no purpose of launching a new visible spectrum space platform unless it is going to seriously dwarf the size of Hubble. Hubble can already be upgraded with new sensors, and has. Which is part of the reason Hubble remains relevant. Hindsight being 20/20 and all, JWST learned from Hubble's need for contact lenses, which would be assumed to follow for whatever next is.

Because servicing it (beyond reboost) may be the same cost as launching something new that offers better performance.

The proposed "better Hubble" that would launch in the 2040s is projected to cost $11 billion. That's a fair few servicing missions; it sounds like we can probably afford to do both, and if so, why not? Keeping Hubble around doing science longer is still useful even with a successor planned.

https://www.nytimes.com/2021/11/04/science/astronomy-decadal...

Not to mention that NASA doesn't have the kind of blank check budget that the military has (by a longshot) — in a budget overview they may have to cut service for old projects to even get research on new projects funded.

when it comes to something like this you are forgetting the government redtape involved in building a new one is likely a bigger cost than the actual engineering building and launching.

Since Hubble is at the upper range of LEO which made shuttle missions difficult, I've always wondered if they could make it slow down enough to lower its orbit more easily accessible for orbiters where part of the refurb is to refuel it so that it could get back to operating orbit.

Maybe not Hubble, but maybe the next gen?

Reply on news.ycombinator.com