VLA on the move
December 30, 2019 11:41 AM Subscribe
The aptly named Very Large Array is an astronomical radio observatory consisting of 27 25-meter dish antennae (plus a spare kept in a shed) laid out on three spokes radiating from a central point. Every four months, the antennae are moved into a different configuration, covering a diameter of 1 km at their closest (Configuration D) to 36 km at their widest (Configuration A). For regular observations, they're mounted on piers sunk 9.7 meters into the ground, but to change configurations, they ride the rails—a double set of train tracks, carried on a custom built transporter. When a dish reaches its new position, the whole antenna + transporter assembly is jacked up, the wheel bogies are rotated 90°, and they then traverse a short rail spur into place (video bookmarked to the interesting part.)
Fascinating - thanks. Big photos of dish antennae like these decorated my childhood home, because my father helped set up the very early NASA tracking network. In the early 60s he'd disappear on business trips to places like Alaska, Hawaii, England, Australia and Ecuador, managing the local engineers who made it all happen.
posted by Rash at 1:23 PM on December 30, 2019 [2 favorites]
posted by Rash at 1:23 PM on December 30, 2019 [2 favorites]
Oh, nice! I've spent many hours at the site, including clambering on the transporter and climbing inside a dish a couple of times. The VLA is *extremely* impressive, even if one is not using it for scientific observations.
posted by RedOrGreen at 2:01 PM on December 30, 2019 [3 favorites]
posted by RedOrGreen at 2:01 PM on December 30, 2019 [3 favorites]
I was there this (soon to be last) year during a work trip and got to climb up into a dish (my second time). It is a really impressive sight!
Ever since one of the dish panels got ripped off an antenna during high winds, it has been deemed too dangerous for visitors to climb all the way up to the dish edge and peer over, unfortunately.
posted by confluency at 2:41 PM on December 30, 2019 [3 favorites]
Ever since one of the dish panels got ripped off an antenna during high winds, it has been deemed too dangerous for visitors to climb all the way up to the dish edge and peer over, unfortunately.
posted by confluency at 2:41 PM on December 30, 2019 [3 favorites]
I have nothing very clever to add to this thread except to share that I once spent about 8 hours broke down on the side of the road in the middle of the VLA (well, the off-center middle that Hwy 60 cuts through), waiting for a tow truck, with no smart devices to distract or inform me (late 80's/early 90's). I... can't really articulate how strange and wondrous a sight it was. A Very Large Array indeed! Mostly clear skies, a few scattered clouds, just New Mexico being stunningly-beautiful, and a lot of large dishes pointing to ... elsewhere in the galaxy.
If you're road-tripping Los Angeles/Santa Fe, I highly recommend using Hwy 60 between Phoenix, AZ and Socorro, NM. It's a gorgeous route and the sudden appearance of the VLA as you dip into the valley is sure to delight any unsuspecting passengers.
posted by ButteryMales at 2:44 PM on December 30, 2019 [4 favorites]
If you're road-tripping Los Angeles/Santa Fe, I highly recommend using Hwy 60 between Phoenix, AZ and Socorro, NM. It's a gorgeous route and the sudden appearance of the VLA as you dip into the valley is sure to delight any unsuspecting passengers.
posted by ButteryMales at 2:44 PM on December 30, 2019 [4 favorites]
I recently visited the VLA myself, which is what inspired this post. Being there made a profound impression on me. Coming from Socorro, we crested a hill 10 miles away and saw it even from that distance.
And I have to admit that I had misunderstood how the VLA worked—I knew that the dishes moved, but I thought they were all on tracks all the time. It makes sense that they would need solid foundations, of course.
There's a plan to expand the VLA from 27 to 244 antennae scattered hundreds of kilometers across the southwest. I'm guessing those won't be mobile.
posted by adamrice at 3:37 PM on December 30, 2019 [2 favorites]
And I have to admit that I had misunderstood how the VLA worked—I knew that the dishes moved, but I thought they were all on tracks all the time. It makes sense that they would need solid foundations, of course.
There's a plan to expand the VLA from 27 to 244 antennae scattered hundreds of kilometers across the southwest. I'm guessing those won't be mobile.
posted by adamrice at 3:37 PM on December 30, 2019 [2 favorites]
As someone who's obsessed with radiotelescopes and other associated megaliths of outer-space infrastructure, visiting the VLA last year was one of the dreams of my life. Seconding everything said above about the strange beauty of the instrument itself and its place in the landscape. And I appreciated that they allow self-guided tours when the official ones aren't up and running.
I'm not usually a gift-shop kind of person, but I dropped $300 that day.
I also appreciated the diminutive Long Wavelength Array just across the street. So easy to miss by comparison, but very cool as well.
posted by mykescipark at 3:39 PM on December 30, 2019 [3 favorites]
I'm not usually a gift-shop kind of person, but I dropped $300 that day.
I also appreciated the diminutive Long Wavelength Array just across the street. So easy to miss by comparison, but very cool as well.
posted by mykescipark at 3:39 PM on December 30, 2019 [3 favorites]
So, I know a fair amount about trains, as I was approaching the VLA for one of their open houses I was very confused when I came across a double-track exempt railroad crossing in the middle of nowhere made me very confused for a few seconds until I looked down the track and thought, "Oh, duh."
(exempt crossings are tracks that are not or barely in use, and double tracks outside of cities and away from buildings usually mean a busy main line that can run trains both directions at once. It was actually the tracks for the transporter, which straddles two tracks.)
posted by ckape at 6:09 PM on December 30, 2019 [3 favorites]
(exempt crossings are tracks that are not or barely in use, and double tracks outside of cities and away from buildings usually mean a busy main line that can run trains both directions at once. It was actually the tracks for the transporter, which straddles two tracks.)
posted by ckape at 6:09 PM on December 30, 2019 [3 favorites]
I was assuming they had some sort of powered locks to secure the dishes to their mounts but nope, plain old nuts and studs put on by a guys with a wrench.
posted by Mitheral at 8:29 PM on December 30, 2019
posted by Mitheral at 8:29 PM on December 30, 2019
Thanks for this post!
I live close by the Westerbork Synthesis Radio Telescope, which isn't nearly as big as the VLA, but still pretty impressive when I visit it occasionally. It has 14 telescopes but only 4 are on rails.
posted by Pendragon at 12:39 AM on December 31, 2019 [1 favorite]
I live close by the Westerbork Synthesis Radio Telescope, which isn't nearly as big as the VLA, but still pretty impressive when I visit it occasionally. It has 14 telescopes but only 4 are on rails.
posted by Pendragon at 12:39 AM on December 31, 2019 [1 favorite]
And I have to admit that I had misunderstood how the VLA worked—I knew that the dishes moved, but I thought they were all on tracks all the time. It makes sense that they would need solid foundations, of course.
I had the same misunderstanding; I'd thought of them as mobile rather than movable.
posted by Ickster at 8:11 AM on December 31, 2019
I had the same misunderstanding; I'd thought of them as mobile rather than movable.
posted by Ickster at 8:11 AM on December 31, 2019
What configuration is the VLA in right now?
posted by peeedro at 8:26 AM on December 31, 2019 [3 favorites]
posted by peeedro at 8:26 AM on December 31, 2019 [3 favorites]
So what are the trade-offs between the different configurations? Why not just always leave them as far apart as possible?
My understanding is that it's like the zoom on a camera, at maximum size it can see more detail but a smaller area.
posted by ckape at 9:26 AM on December 31, 2019 [1 favorite]
My understanding is that it's like the zoom on a camera, at maximum size it can see more detail but a smaller area.
posted by ckape at 9:26 AM on December 31, 2019 [1 favorite]
> what are the trade-offs between the different configurations?
Oh gosh. Fat pitch down the middle.
Okay, at a very basic level: the sensitivity of your telescope depends on its collecting area - how big is the bucket with which you're collecting photons? For the VLA, collecting area is always the same, 27 dishes of 25 m diameter each. That's 27*pi*(25^2 / 4) sq m, not that far off from the collecting area of the 100-m Green Bank telescope (GBT), pi*(100^2 / 4) sq m.
Meanwhile, the resolution - the finest detail a telescope can see - depends on its diameter. For the entire VLA, working as an array, that means the longest baseline, which can change between 1 - 3 - 10 - 36 km.
So different array configurations give you different resolution - and if that's all there was, yes, you'd want to leave it in the maximum array configuration all the time. (Although that's going to be a LOT of data at fine detail, in many cases not so interesting - but whatever, computers, wave hands that way.)
But there are two more things to consider:
How big is the field of view? For a single dish like the GBT, the resolution and the field of view are the same thing - they map the sky pixel by pixel, by pointing at a position and measuring its brightness at a given radio frequency, then the next pixel, then the next ... With an array, though, we can do better! The resolution of our array element - its field of view - is given by the individual dish diameter of 25 m. So the VLA dishes can sample a patch of sky that's (100^2)/(25^2)=16 times larger than the GBT can sample. But that sampled patch can be imaged with a resolution corresponding to the maximum baseline length - so at 36 km, mapping the full field of view will require 36 km / 25 m = 1440 mapping elements - usually ~3 pixels per element - so we're talking about a 4096 x 4096 pixel image grid instead of the GBT's 1 pixel. Again - a lot of computing and data, in fact impossible when the VLA was built in the 80s. (But they could see the future of computing.)
And finally, there's a question about what's the *largest* scale feature the VLA can see. If you're like me and mainly interested in compact sources, we want full resolution all the time. But some people care about extended wispy features, like galaxies, or spiral arms, or molecular clouds - you know, the rest of astronomy. And the most extended feature the VLA can map depends on the shortest baseline.
When the VLA was built, it was hard to conceive of enough computing power to build an instrument that could simultaneously have long baselines for compact sources (36 km) and short enough baselines for extended features. So it was conceived as a multi-scale instrument, where you'd come back to the telescope every 4 months and use the next configuration to fill in details, from the finest features to the most extended features.
Oh, and there's one more trick. What's the shortest possible baseline? It's zero, corresponding to the DC value, the total power, which an interferometer traditionally can't sample. (Even if the VLA dishes touch - and they almost do, in D-array - that's still a >25-m separation.) We can interpolate, but ideally, the zero-spacing baseline is sampled by using a single dish telescope of matched sensitivity - like, yeah, the GBT.
In a future next-generation VLA, we'll just deploy enough small dishes so that a full range of baselines can be sampled at once, from very short to moderately long, giving a wide-field, high resolution image that is also sensitive to extended image features. The computation requirements are - well, kinda' hilarious, honestly. We're struggling right now with EVLA data for the VLA Sky Survey and deploying on national supercomputing facilities. I can't even contemplate the ngVLA.
posted by RedOrGreen at 10:50 AM on December 31, 2019 [17 favorites]
Oh gosh. Fat pitch down the middle.
Okay, at a very basic level: the sensitivity of your telescope depends on its collecting area - how big is the bucket with which you're collecting photons? For the VLA, collecting area is always the same, 27 dishes of 25 m diameter each. That's 27*pi*(25^2 / 4) sq m, not that far off from the collecting area of the 100-m Green Bank telescope (GBT), pi*(100^2 / 4) sq m.
Meanwhile, the resolution - the finest detail a telescope can see - depends on its diameter. For the entire VLA, working as an array, that means the longest baseline, which can change between 1 - 3 - 10 - 36 km.
So different array configurations give you different resolution - and if that's all there was, yes, you'd want to leave it in the maximum array configuration all the time. (Although that's going to be a LOT of data at fine detail, in many cases not so interesting - but whatever, computers, wave hands that way.)
But there are two more things to consider:
How big is the field of view? For a single dish like the GBT, the resolution and the field of view are the same thing - they map the sky pixel by pixel, by pointing at a position and measuring its brightness at a given radio frequency, then the next pixel, then the next ... With an array, though, we can do better! The resolution of our array element - its field of view - is given by the individual dish diameter of 25 m. So the VLA dishes can sample a patch of sky that's (100^2)/(25^2)=16 times larger than the GBT can sample. But that sampled patch can be imaged with a resolution corresponding to the maximum baseline length - so at 36 km, mapping the full field of view will require 36 km / 25 m = 1440 mapping elements - usually ~3 pixels per element - so we're talking about a 4096 x 4096 pixel image grid instead of the GBT's 1 pixel. Again - a lot of computing and data, in fact impossible when the VLA was built in the 80s. (But they could see the future of computing.)
And finally, there's a question about what's the *largest* scale feature the VLA can see. If you're like me and mainly interested in compact sources, we want full resolution all the time. But some people care about extended wispy features, like galaxies, or spiral arms, or molecular clouds - you know, the rest of astronomy. And the most extended feature the VLA can map depends on the shortest baseline.
When the VLA was built, it was hard to conceive of enough computing power to build an instrument that could simultaneously have long baselines for compact sources (36 km) and short enough baselines for extended features. So it was conceived as a multi-scale instrument, where you'd come back to the telescope every 4 months and use the next configuration to fill in details, from the finest features to the most extended features.
Oh, and there's one more trick. What's the shortest possible baseline? It's zero, corresponding to the DC value, the total power, which an interferometer traditionally can't sample. (Even if the VLA dishes touch - and they almost do, in D-array - that's still a >25-m separation.) We can interpolate, but ideally, the zero-spacing baseline is sampled by using a single dish telescope of matched sensitivity - like, yeah, the GBT.
In a future next-generation VLA, we'll just deploy enough small dishes so that a full range of baselines can be sampled at once, from very short to moderately long, giving a wide-field, high resolution image that is also sensitive to extended image features. The computation requirements are - well, kinda' hilarious, honestly. We're struggling right now with EVLA data for the VLA Sky Survey and deploying on national supercomputing facilities. I can't even contemplate the ngVLA.
posted by RedOrGreen at 10:50 AM on December 31, 2019 [17 favorites]
Boy, did this post confuse me at first. The ALMA array is in the thin, dry air of the high Chilean desert, yet the video showed workers chugging away on the mounts, on rail trucks instead of rubber tire vehicles. Then right away the first several comments here were from people who had been to this impossible remote location in South America. Whaaaaat? None of this matches my own knowledge!
OH. The VLA, not ALMA. Carry on :)
(also not to be confused with the VLBA)
posted by intermod at 5:39 AM on January 2, 2020
OH. The VLA, not ALMA. Carry on :)
(also not to be confused with the VLBA)
posted by intermod at 5:39 AM on January 2, 2020
« Older Journalistic crystal ball gazing | Spoilers: 7 Wonders & Scythe Newer »
This thread has been archived and is closed to new comments
Also not that far from Pie Town.
posted by blakewest at 1:19 PM on December 30, 2019 [5 favorites]