Casting a $20M Mirror for the World’s Largest Telescope [video] | Hacker News Reader

Casting a $20M Mirror for the World’s Largest Telescope [video]

spectrum.ieee.org

236 pointssohkamyung54 comments

Fun fact, this facility is underneath the football stadium at Arizona University.

Tom Scott recently posted a guest video from the Active Galactic Videos YouTube channel about this process.

https://www.youtube.com/watch?v=BP9HNVuGb-g

And if you're ever in Tuscon you can visit the facility, they do tours: https://mirrorlab.arizona.edu/tours (Also recommend visiting Kitt Peak and doing one of their night observing programs.)

Ditto that. I haven't done the Mirror Lab tour yet, but have heard great things. Kitt Peak is worth visiting for the night observing programs (they have a couple, see https://www.noao.edu/kpvc/Prog/) as well as just to go up during the day and take in the view.

The Kitt Peak program I did was a great experience. It's in the middle of a desert wilderness, so get there early for incredible views and a sunset. We got a tour of many of the telescopes. Bring lots of warm clothes because it is quite cold and windy after the sun goes down. And then at the end, for one more bit of excitement, you drive down the (steep, winding) mountain road without headlights--just following the brake lights of the person in front of you.

UofA also has a telescope facility on Mount Lemmon with night observations open to the public. The program is equally good.

University of Arizona, U of A or just UA

That's really interesting. Back in my college days, the lab next to me would run experiments that would routinely pickup vibrations from the road nearby. Even when stabilized on an actively controlled optical table. Which isn't that big of a deal, their experiments would run for a few hours. So they would run it in the middle of the night.

The telescope lab is directly under a stadium, their process for a single mirror is several months/years long. I wonder how they manage their process. Do they just pause and continue once the people leave? I also wonder how damaging the starts and stops (if they do it) are to the final quality of the product.

According to youtube tutorials on telescope mirror polishing, a certain amount of randomness in the system is actually quite beneficial. No idea about the final stages, however.

I don't think there is a place called "Arizona University." JshWright means University of Arizona.

I have always place diamond turning and aerospace composites part/tooling very high on my list of high stakes manufacturing.

The process that this mirror goes through, just redefined high stakes for me.

Think about it, if somebody or something makes an error with any significance. Bolt wasn't secured, sensor calibration was off, etc. The entire project is derailed by years. Just getting new material, will take at least a year based on the video.

In aerospace, usually it's alot of money to scrap a part. But then you call up your supplier, and they can make another one "quickly".

With this, you have no wiggle. At all.

I'm curious how they manage risk/failure modes on their team/facility/machines with a project so critical.

Especially since the entire project has a QTY output of less than 10 over a few years.

An earthquake could destroy everything - not only during construction but after installation in Chile as well.

Still, parking the telescope in Chile is easier to fix than Hubble (which has its own famous repair story).

You think the manufacturing is high stakes?

They ship all the equipment for telescopes together on the same ship, because if you lose the mirror you're basically fucked anyway. You're at the mercy of the interstate, the shiphyards, the ship, the seas, the canals, the other shipyards, the other highways, the roads, the drivers, and a bunch of other things.

Manufacturing is important, but it's a lab environment. Getting the stuff where it needs to go is the messy bit.

This even goes into the design. The reason all the telescopes above La Serena are curiously near 8.4M wide is because there's a tunnel which is about 9M wide.

Or like the Hubble, where, effectively, a single area of chipped paint on part of the mirror grinding toolchain almost caused the entire 20+ years in the making project to fail.

My workplace includes a small optics shop where they polish optics that are roughly 2 or 3 inches in diameter. Even in such a modest little shop, the process is just fascinating. The whole shop runs in slow motion, and there is a high level of skill and judgment necessary to make and test the optics while at the same time maintaining the condition of the abrasives and tooling.

Optical manufacturing processes are a wonderful mixture of old and new. Many of the basic techniques are centuries old, but have been updated with sophisticated laser measurements and computer controls.

The timescales involved in the manufacture of this mirror are amazing. I love the way this guy just casually states things like:

"we then cool the glass ... for a few months"

"the mirror will sit for the next year while we machine the back surface flat"

"that final polishing process ... takes about 18 months"

It's kind of mind blowing for it to take so long to make a single object.

There were probably even longer timespans involved in planning and getting it approved by everyone.

I'm curious if its actually polishing for 18 months, or if its a few days of polishing, followed by a few days inspection, on and off.* I can't imagine they simply turn it on and let it run for a year and a half non-stop.

* Or some other tempo of alternation.

>"that final polishing process ... takes about 18 months"

That's a lot faster than it used to be. The 200 inch mirror at Mt Palomar took 13 years to grind and polish. I guess they have greatly improved the technology.

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

And here I'm sitting with my 8" Schmidt Camera and 12" Newton … the amount of engineering that goes into this is incredible. It's not just the mirror, which is of course important, but also having a mount that can precisely track at the sidereal rate with all the mirrors on it. And then, judging by the image, they'll also do Laser Adaptive Optics to correct for atmospheric seeing artifacts, which means adjusting the mirrors and/or camera at a very high rate and /very/ precisely. It basically works by painting and artificial star in the sky with a high powered laser, and by measuring the difference between where it should be and where it actually is, you can then adjust the optical train to compensate. And, then there's the issue with temperature compensation, collimation (looks like a Gregorian-type design), and of course the cameras themselves. For the cameras I'm assuming the illuminated field is quite large, so either this means a HUGE CCD, or stapling several together with all sorts of noise artifacts involved. Very exciting. Very hard.

For those wondering why do they have to go to Chile?, I presume it's for the same reason as given here for the ELT (extremely large telescope, to be completed in 2024): The location was chosen because of its high percentage of cloudless nights and low light pollution [1] and (notably but with a lower weight of importance) the temperature is cold enough that there will not be differences in temperature between the interior and exterior of the telescope which could cause the rays of light to defract (ever so) slightly [2].

[1] https://newatlas.com/extremely-large-telescope-construction-... [2] https://astronomy.stackexchange.com/a/21413

My Mrs is an astrophysicist and has been to Chile about 10 or so times over the last 8 years. She says the Atacama desert is ideal because of Chiles political stability as well as stability of the atmosphere. A high and dry western coastline is good for great “seeing” conditions because highly stable sea air is less prone to fluctuations in temperature which is the bane of any astronomer (this is what creates twinkling of stars). As it happens she was working on the data reduction pipeline for an instrument with adaptive optics (AO) on the Gemini South telescope. AO helps reduce atmospheric disturbances even further so they are now getting better results than Hubble from a ground based telescope.

The same goes for South California mountains, especially if you're into solar imaging (the Big Bear Solar Observatory is there for a reason).

So if a crack did develop in the glass as it was cooling or some other flaw appeared that would make it unusable, could the glass be broken up and be remelted or has its chemistry changed such that it won't meet its specs any longer?

Found a longer video on Youtube. https://www.youtube.com/watch?v=2GufF3pfIyA

Here's another video about the mirror's construction: https://www.youtube.com/watch?v=BP9HNVuGb-g

Given that we're talking about telescopes under construction, the largest one will be EELT, no?

You’re talking about the Extremely Large Telescope, right? What’s the first “E” stand for?

European IIRC.

You are right. Based on an image in this article [1], the EELT would be the largest optical telescope, once completed in 2022, two years after the GMT, if the dates are correct.

And in the battle for superlatives, the cancelled OLT would have been the largest of them all.

[1] https://medium.com/starts-with-a-bang/worlds-largest-telesco...

The battle for superlatives. Off topic, but I have to say that's such a beautiful way to put it. Masterful use of words.

I love the way telescopes are named.

Can't wait for the "JFCTABT".

The GMT FAQ [1] has 2023 for a partially complete telescope (3-4 of 7 mirrors) - only the future will tell if there will even be a point of time where the GMT actually is the largest (optical) telescope in the world.

And for certain applications the four VLT unit telescopes [2] can already be used as one big telescope [3][4]. Even with 4 mirrors in place the GMT would only have slightly more collection area than the four VLT telescopes combined (8.4m primary mirrors versus 8.2m primary mirrors).

[1] http://www.gmto.org/overview/frequently-asked-questions/ [2] https://www.eso.org/sci/facilities/paranal/telescopes/ut.htm... [3] https://www.eso.org/sci/facilities/paranal/telescopes/vlti.h... [4] https://www.eso.org/sci/facilities/develop/instruments/espre...

It would be good to know what off-axis or on-axis means in the context of telescope mirrors.

https://en.wikipedia.org/wiki/Reflecting_telescope#Off-axis_...

1nm precision of the curve and 20nm for surface irregularities. That's incredible. Wouldn't just thermal expansion cause it to deviate from the 1nm precision? What about dust, precipitation just bacteria creating a film on the surface.

Yes, the "nanometer level" is local flatness not the position. The mirror itself is only polished ~20nm flat. It is temperature controlled and made of a low expansion ceramic (zerodur?), but still bent by active optics to make the focus of "guide star" perfect (and to deal with any high altitude atmospheric "seeing" artifacts).

It is temperature controlled and made of a low expansion ceramic (zerodur?) [...]

Borosilicate glass. I also don't think they control the temperatures of the mirrors but I am not sure about that.

[...] but still bent by active optics to make the focus of "guide star" perfect (and to deal with any high altitude atmospheric "seeing" artifacts).

For the GMT atmospheric compensation happens at the secondary mirrors. The primary mirrors are still active to maintain the optical alignment but slowly - less than 1 Hz - as compared to atmospheric compensation in the kHz range and I guess this also compensates for thermal expansion.

> I also don't think they control the temperatures of the mirrors but I am not sure about that.

I know they typically air-condition telescopes during the day to keep the mirror at the anticipated nighttime air temperature, to avoid convection currents caused by a warm mirror. That requirement would seem to preclude controlling the temperature to a specific value, I think.

If anyone is interested in reading about a large telescope project from start to finish, I highly recommend this book: "The Perfect Machine: Building the Palomar Telescope". The project started in the 1920s, but the Palomar telescope (now called the Hale telescope) is still in active use.

If anyone knows of a comparably comprehensive account of a more modern telescope, I would love to hear about it.

> more than 17 kilograms of special glass

That's not very much glass for such a large area, that would make it just a very thin layer... is this correct?

The Youtube video says 17,481.45 kg of glass. Maybe the IEEE article meant to say 17 metric tons.

I never understood why , and . where used as separators... They look so similar especially in hand writing.

I would propose 17'481.45 kg.

That would be the weirdest reason to move to Switzerland or Liechtenstein.

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

Holy shit, this is awesome.

Once these mirrors are mounted and put to work in the telescope, what's the process for cleaning dust off the things?

Each mirror is actually coated on site and periodically when the performance gets bad enough the mirror is removed and taken back to the coating facility where it's cleaned, which actually removes all the aluminium coating the mirror then it's recoated and put back into service.

https://www.space.com/37730-how-to-clean-very-large-telescop...

Imagine having to invent new technology just to be able to create a module for your next programming project.

So just a normal day for a JS developer then?

Why does it matter if the raw glass has imperfections if it's going to be melted?

An imperfection in that would mean there was some problem with the creation of the glass probably some impurity which would cause issues with the thermal expansion properties. If the mirror doesn't expand evenly or as expected it'd ruin the focus which would make the mirror pretty much useless.

Apparently this telescope will have a focal length of 18,200 mm (18 meters).

That's pretty big.

I really like this group photo

http://www.gmto.org/gallery/mirror-lab/#/lightbox&slide=48

From here

http://www.gmto.org/gallery/mirror-lab/

A great photo. Here's the not-quite-so-big (6.5m) multi-element JWST mirror: https://www.flickr.com/photos/nasawebbtelescope/26602405532/...

My brother made some of the instrumentation for one of the earlier telescopes in Chile at the NOAO in Tucson.

Reply on news.ycombinator.com