To my surprise and elation, the Webb Space Telescope is going to work
arstechnica.com
arstechnica.com
I wonder if a future evolution of space telescopes will be some kind of interferometric (obviously extremely hard for IR or visible light, but easier for radio and microwaves) swarm of cheap semi-disposible telescopes than one enormous one.
Then you can add to the swarm, upgrade elements, and retire failed elements without having to eat a multi-billion helping of humble pie.
And rather than have a fearsomely complex integrated sunshield, you could have a similar swarm of simpler satellites that provide a large cool area at L2, and then the observers just need to handle their own heat.
I suppose this could be described as microservices...in spaaaaace. Draw what parallels you will from that!
Producing novel observation platforms is going to be expensive regardless of which platform, because of launch costs.
> I wonder if a future evolution of space telescopes will be some kind of interferometric (obviously extremely hard for IR or visible light, but easier for radio and microwaves) swarm of cheap semi-disposible telescopes than one enormous one.
With space recovery and repair mission capabilities like https://nexis.gsfc.nasa.gov/osam-1.html growing, a single large asset could be easier to repair than constant refresh.
Anyways, small satellites = small sensors.
There is a big difference between a projected life of 1-3 years, and a prestige mission with a long lifespan. Hubble has been operating for ~30 years now.
> Then you can add to the swarm, upgrade elements, and retire failed elements without having to eat a multi-billion helping of humble pie.
Keeping a swarm of space assets ideally situated over time, with proper attitudes and control is a non-trivial problem. Look at Magnetospheric Multiscale Mission, that is considered a hard problem.
Software solutions don't always translate to hardware.
Right, and the GPS constellation has been active for almost thirty years too. I think the GP is probably right that it would be great if the risk associated with a single mega launch/deployment could be spread over a fleet, but ultimately it does no good if you can't use that configuration to get the observations you want, hence the question. It has less to do with it being a "software solution" and more just whether it would actually practically work.
To their point, though, lots of ground-based radio telescopes are now also arrays of many dishes [1], so it's not at all hard to imagine that a similar configuration could be of value in space.
Radio is a billion times easier (at least in naive terms): wavelengths in the kilometers rather than hundreds of nanometers
And yes, they have small elements, that's the whole point. A small element is disproportionately easier to produce than a bigger one. The point of The Swarm would be to offset the small area individually small elements with a large number of them. ALMA does this, and they even pick the dishes up with giant forklifts and move them around to reconfigure the array.
Optical interferometric arrays are very hard and only recently even possible, so it's unlikely you could do it in space "soon" (even LISA is still a long way out, and that's been planned since I was at school). I obviously don't have a handle on if the added noise from positioning errors outweighs the literally astronomical baseline advantage.
And yes, Hubble might have lasted 30 years, but JWST has about 10 years life, and then it's dead and will fall away from L2 unless they can get a refuel/regas mission launched (it does have the ports for it) to it in time.
Of course any number of practical issues can torpedo such a thing from the phase space of feasible implementations, but they're still fun to think about.
This is exactly the point: launch costs are coming down, and there is widespread anticipation that they will start to plummet as the next generation of launchers comes online. If/when Starship (a) becomes available + (b) has any significant competition, launch costs could plummet by an incredible factor.
Is there anything other than manoeuvring/orbit-correction propellant that puts an upper limit of t5he James Webbs projected life? (At least until micro meteors trash too much of the optics?) Is it possible a resupply mission capability could effectivel7y extend its useful life indefinitely? (Did they even bother putting a fuel filler on it?)
Though resupply missions are not yet going out to those orbits, it is completely conceivable that they will in time. For the price of a $100M mission they'll be able to recoup or fix more expensive ones. I normally hear about batteries, reaction wheels, and fuel limiting life, but I think that would relate most to orbit, propulsion type, etc. IDK about a fuel filler.
Reaction wheels (which use momentum to help you point the vehicle) are a semi common failire. There's usually some redundancy built in, but if enough of them wear out you'll lose stability and pointing accuracy.
Even radiation-hardened electronics will wear out over time. If you look at unprocessed images from the Hubble, there's a shitload of pixels that have been fried in the cameras (the effect kind of look like stuck pixels in a computer monitor)
in recognition of getting old(er), i wonder how many youngins even know there's a reference to be caught here.
On duckduckgo I only found some references to The Muppets, a Springer article and a streaming playlist on archive.org
telescopes in spaaaace is so much more than lowly terrestrial scopes. i mean, we don't say telescopes on earrrrrth.
I thought the comment was referring to Portal 2, but apparently it was really referencing pigs in space, and now I learn it's 50s camp?
But the entropy of a large system is low, because it's physically attached with the Strong Force of physics. There is a low risk to other neighbouring satellites (space junk collisions). A swarm of small satellites has high entropy, and is loosely-coupled with the Gravity force of physics.
At what point do we want to accept the financial tradeoffs involved? Humans and the economy also benefit from the large projects, and the management structures also teach efficiency to more people who can go on to create other exciting new sensors.
We could start with a small satellite like Sputnik, and make them grow. Eventually it will reach a point of stability with the neighbouring environment. That could be much larger than we expect. "That's no moon, it's a spaceship!"
I'm going to go and assume that you used the strong force and gravity as metaphors here.
An L2 swarm isn't bound together by gravity, it actually has to maintain active control to stay there, so it's a "self cleaning" area.
Annother option with some of these cheaper launch vehicles is to build the telescope into the upper stage, using it as a bus/platform. So for example you could convert a Starship upper stage into a telescope, using its full 9m diameter for a mirror. Fully assembled on the ground before launch.
That's fabulously expensive: you could have four LHCs, have enough change to replace Arecibo with 10 FASTs and then maybe a aircraft supercarrier on top just for funsies.
It feels like JWST again: it's so expensive that it cannot be allowed to fail, so it's engineered to near perfection, but that costs so much that it has to be engineered even more to be sure, and so on until it's 1900% over budget, a decade late and muscling other science out of funding.
But, also dayum it's a piece of art.
The 2nd mission with which I got involved, an Air Force Captain explained to me that continual trend was to massively underbudget a mission, and then make it whole through cost overruns.
Of course, my work was all about increasing the security capabilities of the constellation, which likely added in the tens, to hundreds of millions in cost to the system.
> It feels like JWST again: it's so expensive that it cannot be allowed to fail, so it's engineered to near perfection, but that costs so much that it has to be engineered even more to be sure, and so on until it's 1900% over budget, a decade late and muscling other science out of funding.
Launch costs are dropping and the space body of knowledge is changing rapidly
System Engineering tools are in their infancy, so rework or changing requirements are pricey. I expect this will really the cost of missions over time, since it directly relates to labor costs.
Simultaneously, the space operator mentality of "extreme availability at all costs" is meeting the hard reality of "space is now an offensive and defensive cyber domain".
We're building too many one-offs and too few constellations, both the birds, and the sensors. If you want lower cost we need to get past unit volume of one.
> muscling other science out of funding
But we don't need four LHCs. We do need a way of observing dim and distant infrared light with high resolution. There's no other instrument that can observe EM radiation from 13 billion years ago. So it's not one or the other; something like the JWST was going to be built sooner or later.
It took a supply shock and a demand shock and trillions for the economy to see inflation, and unfortunately almost all of that is just due to fucking fertilizer and lack of livable wage for logistics workers, not science equipment manufacturers hiking prices...
These mega project that cost a gigantic amount and are hilariously oversubscribed are not actually the best return on investment in my option.
Could you explain what I'm missing? Have we reached a point where the returns of increasing size are very low or is it more that there are enough big mysteries to solve at the smaller size so it isn't much worth increasing until it can be done cheaper?
Say for JWST, there are 18 segments. Could you have 1800 mini Webbs and offset the inaccuracies that come from that with the vastly higher total area? Does that eventually reach an asymptote where a adding more units isn't useful? Does it reach a point where the data re-integration becomes the intractable issue? Can it ever be as good as a gigantic monolith even with an arbitrary number of units? Would the cheapness of individual units be outweighed by the number of them needed?
Combine a wealth of data sources in many different directions and many different observations can also lead to new insides.
When majority of the community, a whole generation of scientists, all have to wait for one massive telescope that can then only do 1/1000 of what people like to do with it, are you really gaining more insight?
Sometimes yes, sometimes no.
[1]: https://www.life-space-mission.com/ [2]: https://lisa.nasa.gov/
According to the simulations performed in [1], an aragoscope of 1km would be able to directly image objects the size of Jupiter's moons (so basically exoplanets even on the small side) 23 light-years away.
EDIT: Took the time to scavenge the numbers: JWST has 0.1 arcseconds angular resolution, a 100m aragoscope would have a 1 milli-arcsecond angular resoulution, a 1km aragoscope 0.1 milli-arcsecond which translates to be able to alpha centauri in a 70x70 pixels patch (and so see its sunspots) or as I mentioned above exoplanets on the smaller scale at 23 light years away (Europa is 1560 km in diameter).
[1] https://www.nasa.gov/sites/default/files/atoms/files/2014_ph...
I'm so pumped to see what science and images the Webb produces.
Could 2022 be the year we find an exoplanet with conclusive biomarkers?
But what other reasons that people might care about X failing as a project when it succeeds as a product?
What should have been done is to have many more missions where these technologies can iteratively been proven so that once you head into a larger project you have some idea of whats gone happen.
And you don't end up launching a telescope that already has 20 year old tech in it.
And the companies that did it made fine profit of course. Promise the moon, don't deliver, make profit. Not a good model.
And how much do those projects cost? And what value do we derive from lots of small iterative solutions?
> And you don't end up launching a telescope that already has 20 year old tech in it.
If you're launching iteratively, then either you're proving out the tech until it's decades old, or you're never making progress on your iterative approach because you're constantly replacing the thing you iterated on with the new hotness.
There is a cogent response to that, as a series of questions.
1 - Which requirements changed?
2 - Which hard science & engineering problems had to be solved, and how trivial, or monumental were they?
3 - Which components failed, or passed testing, requiring rework, or re-engineering?
> Selling a program to taxpayers as a 500 million dollar endeavor and then extracting ten billion dollars from them is the kind of thing that should put people in prison for life.
Initial estimated costs were higher than 500 million. The 500 million number was an NGST estimate, right? I don't think lifecycle costs were ever estimated at 500M, it seems crazy to be that low. Are you sure you are correct on the type of costs you are providing?
Would you say the same about LIGO, for example? Originally thought to be easy, it turned out to be a 40 years long endeavor.
And you are posting this on a forum dominated by software engineers?
Life in prison for being over budget...
Sometimes it's good to overspend if it's an excuse to fund research that would otherwise struggle to find money to stay alive.
Plus I think the budget didn't jump from 500 million to ten billion dollars in on day, project grew and budget grew with it, and someone had to approve it, we are talking about multinational project, so I think everything is well documented and well approved from people who knows the project well better than us.
I do agree with you that if that was military budget, I would think that something is really wrong.
On another article today, again the top response to the top comment was so negative that dang stepped in, and yet again, it's the first thing you see.
It's frustrating, because HN is doing a great job keeping things pretty constructive, but this still seems to reward negativity by thrusting it into view. Shouldn't it at least be below less-downvoted responses?
In 2019, NASA's Astrophysics Division finished up two years assessing the feasibility of assembling a large-aperture observatory in-space. The In-Space Astronomical Telescope (iSAT) Assembly Design Study [1] concluded that In-Space Assembly (ISA) is the only option for building observatories with aperture diameters over 15 m and would still likely be strongly beneficial for smaller ones like the JWST (6.5 m aperture diameter). Efforts like Northrop Grumman’s successful Mission Extension Vehicles, the upcoming DARPA RSGS and NASA OSAM-1 missions, and the usage of Canadarm2 to install instruments with standardized interfaces on the outside of the ISS all demonstrate the increasing maturity of robotic servicing and assembly. The iSAT study describes a telescope composed of modules with standardized interfaces, launched with a spacecraft bus that has attached Canadarm2-like robotic arms that can assemble and deploy modules delivered by space tug from multiple launches. The benefits over launching monolithic spacecraft with hundreds of single points of failure (cough JWST cough) are clear: the mission won’t be limited by a single launch vehicle’s lift ability or fairing size; the same inchworming robotic arm that does initial ISA can later perform repairs and upgrades, either with freshly delivered replacement modules or by debugging malfunctioning parts (see Mars Insight); the final deployed structure doesn’t need to be designed to handle harsh launch conditions; and, design and development will be faster without needing to design and test super reliable deployment mechanisms—if a part fails during orbital checkout, launch a replacement. The primary challenge is designing hardware that today’s limited-dexterity robotics can manipulate, and figuring out supervised autonomy with fallback telerobotics for bringing humans into the loop when needed. There are definitely challenges, but this feels like the right approach. If you could do it near a crewed station for infrequent debugging EVAs, even better. After it's assembled, raise the orbit to L2 with solar electric propulsion.
[1] https://exoplanets.nasa.gov/exep/technology/in-space-assembl...
I'll be writing about this more in Orbital Index (https://orbitalindex.com) sometime soon.
It would be good to develop that capability, but maybe do some trial runs on assembling something a little smaller, and less critical?
[1] Human orbital assembly of the JWST is not possible, because we do not have any crewed vehicles that can make the trip.
I was assuming that GP meant "assembly in Earth orbit" and then the JWST could then rocket off on its own, fully-assembled. (Or, if they didn't mean it that way, I mean it that way.)
I am no expert, but this has to be quite a challenge. How much fuel do you need to do that?
"Neutral Buoyancy Evaluation of Extravehicular Activity Assembly of a Large Precision Reflector" (https://arc.aiaa.org/doi/10.2514/3.26480)
> The procedure and associated hardware are verified in simulated 0-g (neutral buoyancy) assembly tests of a 14-m-diam precision reflector mockup. The test article represents a precision reflector having a reflective surface that is segmented into 37 individual panels. The panels are supported on a doubly curved tetrahedral truss consisting of 315 struts. The entire truss and seven reflector panels were assembled in 3 h and 7 min by two pressure-suited test subjects.
Do you have evidence of this? Assembling things in space would be a major engineering advance with proportional considerations.
No budget to try wildly new things. (The folding was already new enough.)
Also the constant scolding by Congress about the budget overrun has the very predictable outcome of fucking up the overall efficiency of the project. It got assessed and reassessed and committee oversighted, and costcutted...
but of course there was nothing to cut, and everyone knew. It was already risky as fuck and debating how much the sunk cost fallacy applied basically became a pastime.
Perhaps Berger should also credit the engineers, team leaders, administrators, etc. who proved Berger wrong, and the American, European, and Canadian citizens who had enough vision to fund it.
This was a lot of money for some very pure science. If it had blown up on the launch pad (a French launch pad, where government employees were presumably paid to be on a tropical island in December) it would have been Exhibit A in the case against government funding of science for the next 50 years.
There is a very small risk of a launchpad failure, but sending up multiple launches only increases those risks. Further, building a second one would have been much cheaper though still very expensive.
https://en.wikipedia.org/wiki/Timeline_of_Mars_Science_Labor...
TLDR: It was not the EDL [edited to add: entry descent and landing] system, it was the fabrication and integration (putting together) of certain actuators that are used in low temperatures at Mars. People at a contractor, and at JPL, were putting in double shifts and they came within a very close margin of getting it together.
Additionally, there were flight software/avionics issues.
But if you're set to miss the launch window by even a week, you have to wait for the next one 26 months later.
Here is a 1 degree wide view of the neighborhood around HD84406 rendered from Gaia EDR3 data:
https://bsrender.io/sample_renderings/hd84406-1deg-m18-ax100...
Not too many other stars around, and none that are brighter than it within almost a half degree radius. If I zoomed in to the 0.0367 degree FOV of Webb's NirCam instrument, we would only see one other very faint ~21 mag star. However, Webb (and Hubble) can see much fainter stars than Gaia.
Edit: the link posted by muds while I was writing this gives the explanation.
There turned out to be two issues with this choice. The first is that Vega has a very unusual spectrum for a star. This means that more normal stars appear to have weird differences in their magnitudes between different colors. But it's not the stars themselves that are weird, it's just a weird choice of zero points!
A more serious issue became apparent when CCDs became more common in the 1970s and 80s. It turns out that Vega is somewhat variable. You can define the zero point of the magnitude system to be the average brightness over a long period of time, but that doesn't really help you on any given night since the equipment needs to be calibrated daily (or more frequently --- temperature and atmospheric changes require re-calibration).
Another source of annoyance here is that Vega is also very bright. This was a benefit in the days of photographic plates. But modern telescopes with CCDs cannot observe such a bright star. It almost immediately. So this makes calibrating the equipment trickier. (You essentially need a two step process where you use a small, specialized device to calibrate against Vega and then measure the flux from a dimmer reference star, and then measure the reference star with your telescope.)
I guess we can't win :)
- Available for observation for a prolonged time
- A star that has just entered its field of view
- Don't want a star in a field that is too crowded
- The star should be bright, but probably not too bright
Your body is built to expect gravity to pull blood down. The circulatory system thus pushes blood up. When in space this creates problems.
Certainly, any protocols based on handshakes would be a non-starter even at lunar distances (384,000 km, 1.28 light seconds), let alone interplanetary ones (20 minutes to Mars, up to 90 minutes to Jupiter).
At interstellar ranges, the problem becomes even more complex, with keys taking years to be received.
There are some preliminary discussions of space-based IP communications. I do suspect that UDP would be preferred over TCP, for obvious reasons.
My understanding is this is absolutely the case with the Mars Rovers which have engineering versions here on Earth such as this one: https://mars.nasa.gov/news/8749/nasa-readies-perseverance-ma...
I'm not sure it "wouldn't have been that much more expensive" though. Building and testing these things takes a lot of effort, time, and thus, money. Even at just 5% of the costs we're talking about almost $500 million (and it would likely be much more than just 5%).
WH40k is a little bit obscure, but the gist of it is that JWST is designed as an instrument of cosmic revelation. The knowledge it will bring us will shape the future of thought.
If you don't have any spiritual life this might seem meaningless...
foone is a retro hardware hacker who goes on these amazing threads on twitter, and is famous for running Doom on various things, including a pregnancy tester (though admittedly by replacing most of the internals), and his "Carthago delenda est" was on how the JWST was a boondoggle.
Some of that webpage is a bit fluffy for the HN audience, but trust me, look underneath and there's a lot of real stuff there.
For example, GRACE measures groundwater (https://grace.jpl.nasa.gov/applications/groundwater/) and has been the main source of information about fast-depleting aquifers in India and California.
Probably the best measurements we have of whole-atmosphere CO2 (as opposed to in-situ point measurements) come from Earth remote sensing (https://ocov2.jpl.nasa.gov) -- you're right though, that is climate-related.
Another one to take note of is MAIA (https://www.jpl.nasa.gov/missions/multi-angle-imager-for-aer...), which will measure PM2.5/PM10 over various cities.
it's special and different. NASA, and all the national peaceful space agencies, are about scientists using national pride to fund a wide range of projects, nearly none of which have direct impact on terrestrial applied science, but have captured the imagination of the public and leadership, and occasionally do provide practical terrestrial spin-offs (fewer than if we'd invested that money into terrestrial projects, but also very different from them).
look at what the dutch and later the english were doing with merchant ship technology in the 1500s-1800s. It wasn't just good business, it was propaganda to build the power of the Dutch and English states, playing critical roles in th e rapid expansion of those countries into world powers.
The US does what it does to prove that it can easily maintain space leadership (let's just pretend we didn't have to depend on Russia/Soyuz or Russian rocket tubes to put people into space). We can afford this (although, of course, people go starving).
The twitter nerd celebrity machine has gone topsy turvy
Realistically speaking, $10B for a world-class scientific instrument that will operate for 20 years without any human intervention, producing image results that we simply could not have created in any other technically accessible way, sounds like a reasonable budget to me. $25B would be "too much" and $50B would justify inventing a time machine and going back to shut the project down before it wasted too much money.
The defense budget for 2022 in USA is $768 billion for just one year. The budget for the entire development of the telescope was 1.351351% of that budget.
Government budgets are well spent when they are used to do things that are impossible, unfeasible or unwise to do by individual people. JWST is a science investment, which is one of the wisest ways to spend public money.
Anyway I’m sure most Americans would’ve been happy with the 10 dollars a year or whatever it took over the time that 10 billion was spent
Also keep in mind that the original estimate for the program was under $3.5B. That's a cost overrun of almost 3x, which is generally a red flag.
Of course, endlessly spiraling costs is bad. People will disagree about what is acceptable here.
We really did it for nothing but pain and suffering for millions of innocent Iraqi people. And the weakening of our own economy.
Worth it?
So $10 billion on Web vs $10 billion split between 5 other missions that didn't happen, for example, might have been the sort of tradeoff that could be analyzed (then or now).
Or $10++++ billion to enforce the provisions of the UN agreement to end the Gulf War vs. the hypothetical concern that Iraq would create more geo-political problems in the future that would ultimately cost more than taking some action now. Also a tradeoff that is hard to evaluate in hindsight, never mind at the time (hard -- but not immune from criticism or evaluation).
I'm not saying that either of the choices, $10 billion on Webb or $10++++ billion on the Iraq war, were the right choices -- just that the comparison isn't particularly useful, IMHO.
This is a false dichotomy based on the premise that the NASA budget couldn't just be changed, with an added assumption that the Gulf War 2 was either pay all the money we did or not do anything at all. Neither of those are true, which is exactly the point of the comparison.
Someone that was sitting around complaining with us was more familiar with military gear and had some numbers on the cost of training the average fighter pilot. They calculated exactly how many pilot trainings it would take to fund all of NSF Astro for a year. That was nearly a decade ago now, so take it with a grain of salt, but it was in single digits iirc.
Why? JWST is pretty damn revolutionary. There's only so much you can see with visible light. You really want old and far away stuff, go infrared.
Elon "Electric Jesus" Musk only does it as a business model. He'll never do important national stuff that go beyond budget efficiency.
edit:
its payload width*
Since then the price of the rocket has gone from like 20%+ to like 2% or less of the cost. But they will still get the science. So all in all, might be the best investment anybody has ever made in terms of a science project.
The Delta 4, Delta 4H and Atlas 5 would all have been fully qualified to fly the JW and likely could all have launched it. Maybe the Atlas 5 is to small but there were certainty rocket in the US that had the required certification from NASA.
Arianespace was the first commercial launch provider.
So sending astronauts to the ISS and launching spy satellites isn't "important national stuff"?
1. This was not a competitively bid launch. Europe provided the budget for launch so they selected their own launch vehicle.
2. Nothing the Ariane 5 did was very special no matter they hype about 'so exact'. Everybody know that was almost certainty happening. It was typical under-promise over deliver. Other certified rockets would have done the same thing.
3. Both SpaceX and ULA could have launched this mission. Both are full certified for first rate, lowest-risk launches by NASA. Please actually go and look at the list of launches and you will see:
SpaceX:
> https://en.wikipedia.org/wiki/List_of_Falcon_9_and_Falcon_He...
ULA:
> https://www.ulalaunch.com/missions
4. Both ULA and SpaceX are launching both NASA and Dod payloads. Both required separate very detailed qualification.
5. SpaceX also has launched German Defense sats because it was cheaper then Ariane. Italy has just booked a SpaceX flight as well.
6. The SpaceX Falcon 9 is certified and has launched humans. I would consider that 'important national stuff'.
7. Falcon 9 based on most statistical models is now considered safer then Ariane 5. Falcon 9 has also already launched more often then Ariane 5. Insurance cost of payload on Falcon 9 is also the lowest in the industry.
8. Ariane 5 was just grounded for 6 month because of issues with the fairing and required government intervention. This actually did cause significant concern for James Webb lukeyl they had one launch before to test the new configuration. In typical Arianespace fashion they hid this error for as long as they could and didn't tell anybody why they are grounded. Great security culture they got, always trying to brush their failure under the carpet and not talk about it.
Fine to have pride but don't try to bring others down to lift yourself up. It makes you look silly.