When you get some new device and it says leave plugged in for 12 hours to charge first before using, but you don't. I have to wonder with JWST I bet there is a desire to try to have a sneak peek. But I guess calibration is automated so no way to bypass rules. And way too much to lose if something goes wrong ~30 years and $10B.
As I understand things, without the calibration there would be 18 images overlaid each other with some shift in between. Each mirror has quite some degree of freedom, so they have to be aligned to each other until you have one sharp picture. Right now you might see each star 18 times in the picture. Probably they have taken some test images to check the operation of the cameras, but thats about it until a good degree of alignment is achieved.
From the movie Contact. This is probably the sentence which I will never forget and I very often think of if I need to justify buying a backup of anything.
Not only is this a precision instrument, the mirrors are guaranteed to be possible to align within a nanometer once deployed, it's going into space on top of a giant controlled explosion.
A large part of the cost is the manufacture, testing and qualification of the parts and assemblies. Things that doesn't really scale at low numbers.
An international team including NASA has plans in development for a follow up to JWST called The Large Ultraviolet Optical Infrared Surveyor (LUVOIR) [1]. There's two versions, LUVOIR-A with a 15m primary mirror and LUVOIR-B with a 8m primary mirror (compared to JWST 6.5m primary mirror) [2]
From the Wiki "LUVOIR would be able to analyze the structure and composition of exoplanet atmospheres and surfaces. It could also detect biosignatures arising from life in the atmosphere of a distant exoplanet."
Launch cost is actually a _significant_ fraction. Even more so, because a lot of the complexity arises from design and systems to get things right on the first try, since launch costs are so prohibitive.
Put another way, if launch costs were cheaper, it would also have been cheaper to design and build the telescope.
JWST is expensive for two reasons. One is the giant heat shield, as it is an infrared telescope. There is no real way around that.
Secondly it is the space in the rocket, the mirror (and everything else) had to be folded to fit into the rocket. Getting this folding mechanism to reliably unfold was the big challenge - if anything had gone wrong, there would have been no recovery.
With its 8m diameter, Starship could house larger telescopes without folding. Considering how cheap a Starship by itself is, the Starship itself could be the permanent housing of a large telescope. It would just need to jettison the tip to expose the telescope inside.
I believe much of the complexity comes from having to fit it in the nose cone of the Ariane 5 rocket. With a much larger launch vehicle, you can make some different design trade offs to maybe reduce that complexity.
I very much doubt you can build a cheap 6.5m gold plated mirror aligned to a standard that it can get clear images of the farthest corners of the universe, but that can also survive a rocket launch.
There's a relatively new video on YouTube by Launch Pad Astronomy named "How James Webb Orbits 'Nothing'" [0] which left me baffled on how they are capable of orbiting "around" L2, which I had been thinking about for some weeks now and never could find a solution to it.
I think Webb is not the first one to do this, but the video makes it clear how crazy the planning of the orbit insertion and staying in orbit must be.
I hope that the NASA could release some details on this.
Some would say it really isnt an orbit. Orbit presumes a single attractive force (gravity or magnetic) from one body while another moving body is held away by momentum. The circular path around an L point is more a confluance of attractive forces from two bodies. The object is not held up by momentum rather than by balanced pulls of multiple attractive forces. If there is no minimum required velocity to maintain altitude, it isnt really an orbit imho.
From a physics point of view, an orbit around a celestial body is just a special case of a more general phenomenon. The explanation is the same, only the numbers giving the spatial distribution of the gravitational potential within the gravity well are different.
> If there is no minimum required velocity to maintain altitude, it isnt really an orbit imho.
If you put, into a gravity well, a body that is initially stationary with respect to the location of the well's minimum potential, but offset from it, it will begin to move on an orbit, under the influence of that gravitational potential.
The JWST orbit is more complicated, because L2 is a potential saddle, not a well, and it is affected by the moon's gravity, but I think one needs to be aware of the points I made above before one can begin to deal with those complications.
That goes down the road of saying that every circular path is an orbit. I'm reminded of the concept that the sun "orbits" around in the milky way. I don't really consider that an orbit either because the dynamics are totally different. I try to use "orbit" in relation to objects moving around a larger mass and/or a barycenter, which is conceptually a location of average mass. A third tiny object is rather in an orbit around one of the the bodies but that which is constrained by the gravity of the second.
> That goes down the road of saying that every circular path is an orbit.
That is a route I explicitly did not take, because arguments from dictionary definitions, if the issue is not specifically lexical, are usually just pedantic ways of avoiding the issue.
>...the dynamics are totally different.
It was a very important scientific discovery that they are, in fact, very similar, explicable from a small set of universal premises. This is a pattern that is observed widely throughout the sciences. A universe in which just the things you have mentioned were totally (or just largely) different would be a universe with entirely different physics than ours!
>...and/or a barycenter...
You seem to be inconsistent here, as, just before, you rejected the notion that the sun orbits around the milky way.
> I don't really consider that... I try to use "orbit" in relation to...
Claiming that it is wrong to use 'orbit' for the motion of the JWST around Earth's L2 on the grounds that this usage is not in your personal dictionary is not much of an argument - it is very similar to, but even weaker than, the sort of argument that you wrongly insinuated I was making! (And you cannot say I am doing the same, as I am defending the common usage among those who work with this issue - most specifically, those who are controlling the JWST.)
Even the metastability of the JWST orbit is not much of a basis for your position, given that the orbits of the planets in the solar system, and of satellites around the earth, are also ultimately unstable.
On the “sun orbits the galaxy”… I’ve read that the galaxy isn’t bound together by some massive black hole but instead by dark matter. The black hole in the center doesn’t have nearly enough mass for things like the sun to orbit it.
So what is the sun and all its neighbors actually orbiting around?
As big as that black hole is, its gravitational field is nowhere near large enough to hold the galaxy together. What everything is orbiting around is the center of mass of the whole system. [1]
'Everything', here, includes that black hole, all the planets of all the stars (which orbit along with their suns as a solar system), and also the dark matter, if there is any. Dark matter has been hypothesized to explain why galaxies generally seem to be rotating too fast to hold together, based on the mass of the matter that we have observed, and it does so by proposing there is more mass than we thought.
There is another hypothesis (MOND) that posits that the anomaly is on account of gravity being slightly different at long ranges. In this view, too, everything is still orbiting around the center of mass - what's different is the size of the gravitational force each body is subject to.
Note that, in general, you cannot calculate the velocities within a galaxy simply by assuming a point mass, equal to that of the whole galaxy, at its center. In a spherically-symmetric situation, only the mass inside of a body's radius counts, as that outside of it pulls in all directions in a way that cancels out (AFAIK, this is so for symmetric disks as well. [2]) In practice, the motion of any body close to our galaxy's center will be dominated by the central black hole, as it is located close to the galactic center.
They orbit each other. It isnt a bunch of stuff in orbit around something, It is a mass of stuff spining in a big homegeneous blob. Our sun feels far more pull from neighbouring stars than it does anything near the center of our galaxy.
Only the first of those missions was to fix the flaw. The others were to install various upgrades. And all this was only possible because Hubble was in LEO. If JWST fails it’s game over.
NP, just wanted to make sure that was accurate. We transmitted STS-82 live so I remember that one pretty clearly, but STS-61 really was the one that made all the difference. Unfortunately we didn't have our software ready at that time, that only happened in 95 and then NASA and NetNitco talked to us just in time to have it all ready to go for STS-82.
Thinking about it: that's actually scary how quick the time has passed, it's like yesterday. I don't think I left the office from about an hour before countdown until it was back on the ground.
Context for the level of accuracy. Quoting from the article:
To work together as a single mirror, the telescope’s 18 primary mirror segments need to match each other to a fraction of a wavelength of light – approximately 50 nanometers. To put this in perspective, if the Webb primary mirror were the size of the United States, each segment would be the size of Texas, and the team would need to line the height of those Texas-sized segments up with each other to an accuracy of about 1.5 inches.
Modern hard disks have a track pitch in tens of nanometers, and the read head has to move with matching precision, it also flies 3 nanometers above the surface.
Wow, that's really cool. Thanks for sharing the knowledge. I didn't realize how physically precise modern hard drives were! I suppose if it makes sense if you think about how much information they are storing in a small space.
Totally naive question here - can all this not be done automatically? It seems to me as if this is a largely manual process, and that's why it's going to take 3 months, no?
I guess you imagine the control room is joy-sticking / flipping each individual switch?
The processes are automatically executed, but are check-pointed rather than run end-to-end. This is to make sure partial failures can be detected.
Building the self-sensing and error-recovery logic to support a "level 5" automated unfolding is wasteful of taxpayer money when a few folks can just take a little longer to do it stepwise -- with much less development, hardware, and software cost overall.
I've worked on a bespoke science experiment not too dissimilar to JWST (though much cheaper). These types of machines never resemble anything like mass produced parts. All of the tooling is custom made for one shot. The next machine will be the same story. What we take forward to the next machine are the lessons learned and technology developed.
It has to be for the same reason that creating a one-off prototype of a new plastic widget involves a much more manual process than a scaled-up assembly line. This is the first time this process will be performed on a real James Webb in space, and between each calibration step, they'll need to verify the results are as modeled in theory and adjust or reiterate if not.
Generally automation is applied to repetitive procedures with well understood bounds.
In this case all but the fine-phasing and possibly FOV alignment are one-off operations during commissioning. The pre-launch simulations may not match the actual behaviour and results of the telescope so it makes sense to perform each step under engineer/scientist control and be able to spend time studying the data and possibly improving the process along the way and revisiting earlier steps and/or segments.
The Keck telescope, which is a ground-based telescope with 36 actively aligned mirrors, is aligned automatically twice a second.[1] The original control system used a VAXstation II, but they've probably upgrade by now. The Keck has a tougher alignment problem - gravity warps the mirrors as the telescope is pointed, so they're constantly making corrections.
The Webb telescope is 10 light-seconds out, so manual adjustment will be really slow. NASA is being very cautious. It can't be serviced, and the US will probably never launch another one.
I would guess it's semiautomatic-- lots of humans squinting at interference patterns, guessing at an adjustment, then waiting 24 hours to get data back.
Mirror is supposed to be at ~50K, and it isn't there yet. (And the instruments are still at 100K+, a month after the sunshield deployed, and you sure aren't doing anything until the actual cameras are cold!) Every time you run an alignment motor you heat up a mirror segment a little bit, so you do want to run coarse alignment early, then gradually converge on fine alignment as the whole structure approaches design temp.
Remember, JWST is weird. Hubble was one big piece of (warm) glass. Multisegment telescopes down on Earth can be nailed to a hundred ton support structure, but JWST is comparatively big and floppy, since the frame had to be light enough to launch into space. (Every kilogram of mirror support structure is another kilogram of propellant lost, and once all the fuel is gone you don't have an observatory anymore...) Something like that is just going to inherently take a while to calibrate.
The temperature of deep space is 2.7K due to the cosmic microwave background. JWST's cold side is going to be about 40K, so it's substantially warmer than that.
I'm not sure how much of that heat is coming through the sunshield and how much is heat from the equipment.
I'm not a physicist I just watched the NASA video explaining that the instruments will need an active cooling system called the cryocooler that cools the stuff below what it can achieve in the vacuum of space on its own.
I think that's a pretty poor way for NASA to have put it, because the vacuum of space 1AU from the Sun isn't particularly cold as such. Actually, as a vacuum, it tends to make things there hotter up to a point, not cooler, because you can only radiate that solar heat away. So satellites that are in sunlight, like you are at L2, can easily be at 150°C and even up to 300 depending on various factors.
If the JWST were out by, say, Pluto, then you certainly could describe that vacuum as "cold" (inasmuch as a vacuum has a temperature), because you hardly get any sunlight. In fact, Pluto itself has a temperature around 40K.
Design temp for the MIRI detectors is 6.7K. The front end electronics produce some heat, and are in a metal box sealed against light leaks, so they'll need some active cooling.
Skimming this paper https://www.stsci.edu/files/live/sites/www/files/home/jwst/d... it looks like they expect 10.46mW optical module heat, and they've given themselves 36mW of headroom "for conductive and radiative loads." (Heat conducted from the structure, and radiant heat from the the comparatively warm primary mirror?)
Pretty sure I read this phase takes so long because the initial alignment is done using the same extremely precise, slow-moving actuators as are used for focusing observations, on a per-mirror basis.
It's not because it's a manual process, it's because they opted for simplicity in the actuators department rather than having another coarse+fast mode in the physical mechanism, or redundant mechanisms.
if you automated it, then what are all of these people going to do? all you automation people wanting to take away the jobs of all these people! this is why unions exist! ;-)
I have to make an orderly retreat to my next cynical line: everyone in that room is aware that their reaction wis filmed and will be shown across the world.
Your team spent years researching something that had never been done before. Survived many rounds of "should we just cancel this project", continued to be funded after coming in no where near your proposed budget, delay after delay, to finally see the thing lift off. You wouldn't have the slightest bit of non-robotic emotion for yourself and fellow team members?
Hope I never have to work with someone as dull as that.
If you're a developer that has automated something that does something faster/cheaper/more accurately than a person and haven't had this thrown at you, then you're missing out! It's like a rite of passage.
> Although Image Stacking puts all the light in one place on the detector, the segments are still acting as 18 small telescopes rather than one big one. The segments need to be lined up with each other with an accuracy smaller than the wavelength of the light.
Why is this necessary? Phase difference shouldn't matter unless the light is coherent, which I wouldn't expect in starlight. Which assumption is wrong?
For interference the light needs to be able to take two separate paths with different length, which is the case in the oil film example or with two slits. In a telescope, each photon can only take a single path. So what would it interfere with? If it's a different photon then coherence seems to be required
That's a classical-world blunder: single photons can and do interfere with themselves.
Making a measurement of which path a photon took, is forcing the photon to have taken one path. If you don't make the measurement (read, if you don't interact with the photon somewhere along some path it could take), you can get interference between the paths, despite it being a single photon.
Yes I understand that: the single photon interference in the double slit experiment. But what two paths could a photon take in a telescope allowing it to interfere with itself? The mirrors are placed so far apart a photon could never bounce off more than one mirror?
Sorry but you didn’t seem to really understand the implications of light being a wave here.
Forget about photons altogether and think of light as a wave and you will understand that yes of course, light bounces on all mirrors and interferes across mirrors. Who cares about the distance between mirrors really, light could (and does) interfere with mirrors miles apart.
Hm? Oil slick iridescence and the original double slit experiment demonstrate entirely classical wave behavior. No QM needed. Same thing with Webb. Light arriving from the same source but in slightly different phases absolutely causes interference patterns explainable in purely classical terms.
When thinking about something like a telescope, you pretty safely ignore the concept of "photons" and think about light entirely as a wave. And waves do interfere with themselves.
Light as something made out of particles is relevant only for the photovoltaic effect and quantum-scale interactions.
Obviously I'm wrong but I'm not sure why. Light bouncing off of two mirrors can interfere constructively or destructively, depending on the phase difference. The phase difference from light bouncing off of two aligned mirrors is random (as it's non coherent light). Adding a constant in the phase difference due to misalignment would still lead to a random difference. So it seems to me the phase shift wouldn't make a difference.
The phase difference between light bouncing off two aligned mirrors is NOT random. This is what you are missing.
Non-coherent light only means it is made of many wavelengths at the same time (Think of it as the sum of many « coherent lights »). Interference does happen even for incoherent light, within each wavelength.
Example: oil iridescence on water is interference which works on « incoherent » light.
