Photons incoming: Webb team begins aligning the telescope

https://blogs.nasa.gov/webb/2022/02/03/photons-incoming-webb-team-begins-aligning-the-telescope/

by kloch • 4 years ago

209 80 4 years ago

blogs.nasa.gov

divbzero• 4 years ago

It’s incredibly exciting to watch Webb deploy. I was too young to remember Hubble but will definitely remember this.

guerrilla• 4 years ago

It makes me feel like a kid again.

fsckboy• 4 years ago

Hubble deployed only to discover they had screwed up and it was blurry.

https://www.nasa.gov/content/hubbles-mirror-flaw

jacquesm3• 4 years ago

Which was then fixed in space by astronauts during 5 consecutive EVAs.

virgildotcodes• 4 years ago

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?

dylan6043• 4 years ago

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! ;-)

BurningFrog3• 4 years ago

That may be overly cynical.

But what I always think when a NASA control room erupts in joy after a landing, is that those people just learned they have a job for years.

bluquark• 4 years ago

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.

iam-TJ• 4 years ago

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.

jvanderbot• 4 years ago

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.

thunderrabbit• 4 years ago

Building the self-sensing and error-recovery logic to support a "level 5" automated unfolding is wasteful

Very true, especially if they tried to do all that the first time.

In a way, this stepwise deployment is a test run for a future version that can do it a little bit more automated.

Animats• 4 years ago

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.

[1] https://www.osti.gov/biblio/6240189-alignment-calibration-ke...

dtgriscom• 4 years ago

Clarification; 5 light seconds from Earth, but 10 light second round-trip.

pengaru• 4 years ago

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.

sbierwagen3• 4 years ago

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.

Also note that JWST still hasn't cooled down to temp yet! Temps here: https://www.jwst.nasa.gov/content/webbLaunch/whereIsWebb.htm...

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.

stavros3• 4 years ago

Wait, how is it still warm? I guess vacuum is a much better insulator than I thought, does it just take way too long to radiate heat?

Gare• 4 years ago

then waiting 24 hours to get data back

Why 24 hours? The JWST is only 5 light-seconds away and has up to 28 Mbit/s downlink.

dghughes• 4 years ago

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.

RosanaAnaDana• 4 years ago

I think its much more than the sunk cost we would lose. There would never be another.

dmitryminkovsky2• 4 years ago

Why not? You mean JWST specifically or some reason there can’t be a similar telescope?

QuadrupleA• 4 years ago

I often wonder with JWST, wouldn't building another cost way less than $10B? We have detailed blueprints, the software, etc.

_ph_2• 4 years ago

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.

rswail• 4 years ago

Watching the Webb unfold has been both heart-stopping and amazing.

Just the heatshield itself is astounding and bodes well for future human travel.

A_No_Name_Mouse4• 4 years ago

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?

sbierwagen3• 4 years ago

You can get interference patterns off of white light: https://www.quora.com/Can-white-light-produce-interference-p...

After all, physicists were producing interference fringes in the 18th century, and they sure didn't have lasers back then: https://en.wikipedia.org/wiki/Young%27s_interference_experim...

A_No_Name_Mouse4• 4 years ago

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

radium3d• 4 years ago

Show us the photons! :)

tomas789• 4 years ago

I would be interested in reading something more in-depth. Especially how the algorithms behind the alignment work.

qwertox2• 4 years ago

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.

[0] https://www.youtube.com/watch?v=ybn8-_QV8Tg

sandworm1013• 4 years ago

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.

mannykannot3• 4 years ago

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.

ColinWright• 4 years ago

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.

dwaltrip2• 4 years ago

Does this mean that they have some sort of actuator that can move with nanometer precision? I didn’t realize that was possible… wow.

GuB-42• 4 years ago

The best part is that you may already have one.

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.

Photons incoming: Webb team begins aligning the telescope

Related Stories