Can telescopes in space replace ground based telescopes?

[envy887]

LSST isn't trying to resolve anything, AFAIK. It's just gathering light. So there's no point in interferometry, and if you can run longer exposures, then there's no need for the huge mirror. Using smaller mirrors and building more of them gives a double helping of cost savings.

Also, ISTM the problem with deep space orbits for a survey constellation is getting all the data back. DSN is slow and already too busy. I think a TESS-like elliptical orbit where download happens near perigee at high data rates solves this nicely.

[Swedish chef]

Take a ring of arbitrary diameter and a few cm thickness. Glue a sheet of Mylar or other suitable material to either side. Mildly pressurize the then enclosed space.

I'm somewhat fond of this idea, to get rid of the expensive and heavy main mirror and launch or perhaps fabricate something of lesser quality, then when that not so ideal mirror is in space use adaptive optics to straighten it out. If you can bend a perfectly ground mirror out shape to overcome the trouble of atmosphere here on earth I'm guessing the same thing could be used in space on a mirror of lesser quality.

Now the adaptive optics on a telescope here on earth are constantly moving to adjust for the atmospheric effects, my suggestion would only have to be actuated once, or every few days to overcome heat issues deforming the main mirror. (I would probably try the adaptive optics on the secondary mirror since they need something sturdy and therefore heavy and expensive to act upon)

Might this work or is there they usual problem of garbage in, garbage out?

[OTV Booster]

LSST isn't trying to resolve anything, AFAIK. It's just gathering light. So there's no point in interferometry, and if you can run longer exposures, then there's no need for the huge mirror. Using smaller mirrors and building more of them gives a double helping of cost savings.

Also, ISTM the problem with deep space orbits for a survey constellation is getting all the data back. DSN is slow and already too busy. I think a TESS-like elliptical orbit where download happens near perigee at high data rates solves this nicely.

I’m not sure one way or the other on the interferometry. The resolution is seeing limited at .7”. Not sure what diameter is needed to get this in the optical range.

Wikipedia has a good overview of both the optics and the data handling. The data flow is incredible. They are budgeting for ONLY 1.28 petabytes/year because there will be down time for maintenance and weather. I don’t know the breakdown on that.

There is nothing out there now or in the foreseeable future that can handle a data pipe that big. That’s where making projections through the disjunction comes in.

In the 10 or 15 years (did you see me wave my arms <g>?) it will take for the next generation of space scopes to show up it’s impossible to know what sort of space coms will be available. Whatever it is, the coming revolution in access to space means we can not do a linear projection.

Future comms could range from a solar system wide StarLink on steroids to something wacky like a signal carrying laser pumping a spot on Jupiter or the suns corona for observation from earth. Don’t spend too much time shooting down that last. It’s a place holder for ‘something completely new that we’ve never thought of.’

The StarLinks that will degrade surface observation will contribute to the revolution that will allow observations we can only dream of now.

Here’s a standing bet: within the next 20 years we will be able to image the Alpha Centauri system in the optical range at the same resolution the 200” can give at Jupiter, and it will be done from space. I’d cut that down to 15 years but predictions are hard, especially if they’re about the future.

Phil

Edit to add: I’d make the bet a beer but if I’m still around in 20 years my doctor probably won’t let me drink it. 

[OTV Booster]

Take a ring of arbitrary diameter and a few cm thickness. Glue a sheet of Mylar or other suitable material to either side. Mildly pressurize the then enclosed space.

I'm somewhat fond of this idea, to get rid of the expensive and heavy main mirror and launch or perhaps fabricate something of lesser quality, then when that not so ideal mirror is in space use adaptive optics to straighten it out. If you can bend a perfectly ground mirror out shape to overcome the trouble of atmosphere here on earth I'm guessing the same thing could be used in space on a mirror of lesser quality.

Now the adaptive optics on a telescope here on earth are constantly moving to adjust for the atmospheric effects, my suggestion would only have to be actuated once, or every few days to overcome heat issues deforming the main mirror. (I would probably try the adaptive optics on the secondary mirror since they need something sturdy and therefore heavy and expensive to act upon)

Might this work or is there they usual problem of garbage in, garbage out?

Adaptive optics are heavy and take a fair amount of power by satellite standards. ISTM that any curvature problems can be dealt with through corrective optics.

Mylar has a great level of smoothness but it’s probably not a fraction of a wavelength. Maybe another material or maybe a microgravity manufacturing technique can fix this.

Any thermal distortion is best handled by avoiding it in the first place. That’s why my first instinct was to station far out and do a stand off heat shield.

It may be that innovative space built mirrors will not be of inferior quality. Once launch becomes relatively inexpensive and infrastructure is built up the scope itself may become much less expensive per meter of aperture with no loss of quality. It will be exciting to watch.

Adaptive optics also work to adjust mirror sag with changing angle. Another problem that goes away with space based scopes.

Phil

[Zed_Noir]

<snip>
Also, ISTM the problem with deep space orbits for a survey constellation is getting all the data back. DSN is slow and already too busy. I think a TESS-like elliptical orbit where download happens near perigee at high data rates solves this nicely.

A wacky idea about transferring survey data back from a deep space location.

Packed a large unmanned space vehicle with a few hundred tonnes of SSDs along with a short range laser communication system to retrieved the data in the vicinity of a constellation data nexus.

There could be a relay of SDD packed data retrieval vehicles to retrieved the data periodically from each nexus.

Of course a few hundred tonnes of high capacity SSDs will not be cheap for even a single data retrieval vehicle. Think you need at least a handful of such vehicles.

[OTV Booster]

<snip>
Also, ISTM the problem with deep space orbits for a survey constellation is getting all the data back. DSN is slow and already too busy. I think a TESS-like elliptical orbit where download happens near perigee at high data rates solves this nicely.

A wacky idea about transferring survey data back from a deep space location.

Packed a large unmanned space vehicle with a few hundred tonnes of SSDs along with a short range laser communication system to retrieved the data in the vicinity of a constellation data nexus.

There could be a relay of SDD packed data retrieval vehicles to retrieved the data periodically from each nexus.

Of course a few hundred tonnes of high capacity SSDs will not be cheap for even a single data retrieval vehicle. Think you need at least a handful of such vehicles.

Woo Hoo. Someone as wacky as me and braver about it too.

I was thinking about a passel of thumb drives brought back my cube sats. Treat each cubesat like a RAID drive and if you miss one, de nada.

The investigators will go nutso waiting for the pipeline to fill.

Phil

[OTV Booster]

<snip>
Also, ISTM the problem with deep space orbits for a survey constellation is getting all the data back. DSN is slow and already too busy. I think a TESS-like elliptical orbit where download happens near perigee at high data rates solves this nicely.

A wacky idea about transferring survey data back from a deep space location.

Packed a large unmanned space vehicle with a few hundred tonnes of SSDs along with a short range laser communication system to retrieved the data in the vicinity of a constellation data nexus.

There could be a relay of SDD packed data retrieval vehicles to retrieved the data periodically from each nexus.

Of course a few hundred tonnes of high capacity SSDs will not be cheap for even a single data retrieval vehicle. Think you need at least a handful of such vehicles.

On a more serious note looking ~15 years down the road we can expect a more robust space comm infrastructure.

DSN will be far surpassed. Here comes the hand waving.

StarLink and it’s competitors will have the capacity to link to anything below GEO. GEO will have ability to link to anything cislunar including L1 halo and Earth-Sun L4&5 which in turn can link to E-S L1. These in turn will be able to link to Mars- Sun L4&5... etc.

That’s probably overkill for solid uninterrupted comms between Earth, Luna and Mars but you get the idea. Comms for an ongoing investigation beyond Mars will need a commitment for an extension of the comms infrastructure or it won’t be feasible.

Using current technology the bandwidth would be inadequate. Extrapolating for 15 years is dicy but my guess is some subset of Moore’s law applies. Far UV or X-ray lasers come to mind along with all the multiplexing and phasing games that comes with multiple channels.

It really is impossible to extrapolate 15 years. My guess is that if space infrastructure does not progress to something along these lines (capability, not specific architecture) in 15 years it will be a sign that Musk’s plan to get all the eggs out of one basket is too little, too late. Mars will never thrive if it’s a communications backwater.

Phil


Edit: changed a couple L2’s to L1’s.

[hop]

A wacky idea about transferring survey data back from a deep space location.

Packed a large unmanned space vehicle with a few hundred tonnes of SSDs along with a short range laser communication system to retrieved the data in the vicinity of a constellation data nexus.

Note that near real-time alerts for transients is a major objective for LSST (see https://www.lsst.org/about/dm/data-products).

AFAIK expected overall data volume from is on the order of ~20TB/night. Not impossible to do from space, but definitely a substantial challenge compared to current missions, especially outside of LEO.

[Robotbeat]

I doubt there is anything that must be done on the ground, but even if launch is free, the cost of reconstructing humanity's collective ground-based capabilities in space would be, er, astronomical.

The cost could actually be *lower* due in part to lower gravity.

Large ground observatories are super expensive. Keck has to be constantly refrigerated to near-freezing during the day so that it's at the same temperature as at night (so as not to distort the telescope). They're enormous objects that need to be precisely pointed.

In space, you don't need a laser guide star. You're not perturbed by day, by clouds, by the Moon, etc. That means you probably have 3 times the observing time. You can observe in any direction. You can observe (fairly) close to the sun.

To assembly a 14m reflector assembly only takes about a 5 hour EVA for a couple astronauts. We could just build a whole bunch of large aperture telescopes in orbit. One crew could build dozens of them in a single 6 month stint. And the weight is lower than deployable structures. Much lower hardware cost, too.

So I actually think you COULD replace ground telescopes with space-based ones.

The smaller scopes would also be needed.

But we need not just cheap launch, but cheap human workers in orbit.

[OTV Booster]

I doubt there is anything that must be done on the ground, but even if launch is free, the cost of reconstructing humanity's collective ground-based capabilities in space would be, er, astronomical.

The cost could actually be *lower* due in part to lower gravity.

Large ground observatories are super expensive. Keck has to be constantly refrigerated to near-freezing during the day so that it's at the same temperature as at night (so as not to distort the telescope). They're enormous objects that need to be precisely pointed.

In space, you don't need a laser guide star. You're not perturbed by day, by clouds, by the Moon, etc. That means you probably have 3 times the observing time. You can observe in any direction. You can observe (fairly) close to the sun.

To assembly a 14m reflector assembly only takes about a 5 hour EVA for a couple astronauts. We could just build a whole bunch of large aperture telescopes in orbit. One crew could build dozens of them in a single 6 month stint. And the weight is lower than deployable structures. Much lower hardware cost, too.

So I actually think you COULD replace ground telescopes with space-based ones.

The smaller scopes would also be needed.

But we need not just cheap launch, but cheap human workers in orbit.

An advantage you missed is the optics would be refraction limited rather than atmospherically limited. For he cost of the time needed to collect more photons, a small scope can do the same job as a large scope.

The cost of human workers is less wages (as I’m sure you realize) than the cost of maintaining infrastructure. As launch costs drop, infrastructure costs drops. The meat bags can pull in $1m/year each and it’s not even a rounding error until launch costs drop to something ridiculous.

Phil

[Proponent]

The cost could actually be *lower* due in part to lower gravity.

Large ground observatories are super expensive. Keck has to be constantly refrigerated to near-freezing during the day so that it's at the same temperature as at night (so as not to distort the telescope). They're enormous objects that need to be precisely pointed.

In space, you don't need a laser guide star. You're not perturbed by day, by clouds, by the Moon, etc. That means you probably have 3 times the observing time. You can observe in any direction. You can observe (fairly) close to the sun.

To assembly a 14m reflector assembly only takes about a 5 hour EVA for a couple astronauts. We could just build a whole bunch of large aperture telescopes in orbit. One crew could build dozens of them in a single 6 month stint. And the weight is lower than deployable structures. Much lower hardware cost, too.

The 39-m ELT has a price tag of about a €1.2 billion euros.  Allowing for some cost growth, that's perhaps one-fifth the cost of the 5.6-m JWST.  JWST employs robotic systems for deploying both mirror and sunshade, but I've never found in the GAO's reports any indication that those systems drive the bulk of the cost (I've looked, because advocates of Ares V and SLS used to claim that JWST would be much cheaper if only its mirror did not require deployment).  JWST's optical tolerances are looser, because it's optimized for the infrared, yet still it's mirror is made of beryllium, for thermal reasons.  I'm pretty sure refrigeration of large ground-based telescopes is far cheaper than JWST's method of refrigeration, namely sending it to Earth L2, with a whopping huge sunshade.

EDIT:  JWST's requirements for cooling are much more stringent than those for ELT, since JWST operates well into the infrared.  Hence, the comparison is not fair.

[su27k]

To assembly a 14m reflector assembly only takes about a 5 hour EVA for a couple astronauts. We could just build a whole bunch of large aperture telescopes in orbit. One crew could build dozens of them in a single 6 month stint. And the weight is lower than deployable structures. Much lower hardware cost, too.

Is this just back of envelope guesstimate, or has this been studied?

I do hope NASA does some study on this subject, given we may have the launch capability for this in the near future.

[trm14]

To assembly a 14m reflector assembly only takes about a 5 hour EVA for a couple astronauts. We could just build a whole bunch of large aperture telescopes in orbit. One crew could build dozens of them in a single 6 month stint. And the weight is lower than deployable structures. Much lower hardware cost, too.

Is this just back of envelope guesstimate, or has this been studied?

I do hope NASA does some study on this subject, given we may have the launch capability for this in the near future.

Yeah, I'd like to see some source for that, too. It sounds quite dubious to me.

[OTV Booster]

To assembly a 14m reflector assembly only takes about a 5 hour EVA for a couple astronauts. We could just build a whole bunch of large aperture telescopes in orbit. One crew could build dozens of them in a single 6 month stint. And the weight is lower than deployable structures. Much lower hardware cost, too.

Is this just back of envelope guesstimate, or has this been studied?

I do hope NASA does some study on this subject, given we may have the launch capability for this in the near future.

Five hours - maybe after getting a few under their belt. This would not include alignment and collimation. If it’s a segmented mirror this could be a slow job. Automated actuators could speed this up but add a lot of mass. They could be part of a tool kit and one set used for multiple mirrors.


Making the machine is hard. Making the machine that makes the machine is really hard. It would boil down to production numbers.


Phil

[eeergo]

Another data point:

https://twitter.com/jaynebirkby/status/1699410725184123149

GMT: 10x Hubble res, 4x JWST res, $2B-class project, NET 2029, decades of continuously upgreadable observations
LUVOIR (proposed): 5x JWST res, $10B-class project, NET 2039, 5-10 years of primary+extended mission

Ratio LUVOIR/GMT: 1.25x resolution, 5x cost, similar (?) development periods, substantially longer operations

[Robotbeat]

It will take a completely different approach to make space telescopes competitive. It’s possible, but it has definitely not yet been done. No one has taken an extremely cost-aggressive approach like Starlink for space telescopes.

(I do not mean to imply you’d use distributed satellites… monolithic has a lot of advantages. Secondly, Starlink has a massive commercial market whereas space telescopes don’t really.)

That said, a factor of 5 difference in cost is much smaller than I expected!

A space telescope can operate for 2-4 times as often (not limited by weather or daylight or seeing conditions), and it can see targets in both hemispheres without problem. It also has full spectrum capability.

[eeergo]

It will take a completely different approach to make space telescopes competitive. It’s possible, but it has definitely not yet been done. No one has taken an extremely cost-aggressive approach like Starlink for space telescopes.

(I do not mean to imply you’d use distributed satellites… monolithic has a lot of advantages. Secondly, Starlink has a massive commercial market whereas space telescopes don’t really.)

That said, a factor of 5 difference in cost is much smaller than I expected!

A space telescope can operate for 2-4 times as often (not limited by weather or daylight or seeing conditions), and it can see targets in both hemispheres without problem. It also has full spectrum capability.

Of course, $10B is the minimally credible budget that can be quoted after JWST - it's likely it will end up costing significantly more, just like JWST was supposed to be $1B in its infancy. Probably not such an extreme overrun in percentage ($0.1T is hardly justifiable) but still. The estimated running cost for those 10 years of ops, by the way, is between $8-14B, or around $1B/year. I didn't quote it because I couldn't find such a figure for GMT, but I'm fairly certain it won't be close to even the same order of magnitude, not even including upgrades or accounting for the extra observing time available.

[Robotbeat]

JWST was massively mismanaged. No one will believe it now that it has successfully launched, but JWST should’ve been canceled a decade or two ago. Its successful launch, normalizing this borderline corrupt incompetence by contractors (and dishonesty by the proposers, who knew it’d cost VASTLY more than they claimed), is a Pyrrhic victory.

[eeergo]

JWST was massively mismanaged. No one will believe it now that it has successfully launched, but JWST should’ve been canceled a decade or two ago. Its successful launch, normalizing this borderline corrupt incompetence by contractors (and dishonesty by the proposers, who knew it’d cost VASTLY more than they claimed), is a Pyrrhic victory.

Mostly every major innovative project in spaceflight (and outside of it, to be honest) should have been cancelled by that metric. Not trying to redeem the incompetence and dishonesty that I agree was avoidable, but it was hardly a one-off, except perhaps for the absolute and relative sizes of the overrun. That's why I made the caveat nobody should expect a 10x cost increase for LUVOIR, but neither the cost estimate made decades before its launch when it's just a concept. The GMT budget has much less play in it, if only because first light is expected in about 5 years and construction is well underway.

[Robotbeat]

Lots of projects were managed far, far better. Kepler, the Mars Exploration Rovers, etc. The most insidious idea is that the mismanagement doesn’t matter because “it costs what it costs.”