The Square Kilometre Array

  • I
  • Thread starter alberto91
  • Start date
15
19

Main Question or Discussion Point

As some of you probably know, the Square Kilometre Array will become the biggest radio telescope on Earth, with a collecting area of 1 square kilometre.

The construction will start in 2021 and the first light is expected to take place in 2027. It will cover the frequencies from 50 MHz to 15 Ghz.

But what I wanted to share with you guys is a new study about how far the SKA can 'listen'.

A recent study points out that the SKA could detect extraterrestrial airport radars 200 light years away.

Source:
What do you guys think?
 

Answers and Replies

berkeman
Mentor
56,131
6,156
Very interesting, thanks for posting. It looked like the lower frequency dipole array uses regularly-spaced antennas, but what is up with the irregular spacing of the higher-frequency parabolic dishes? Is there a reason for the apparent random spacing shown in the simulation/rendition video?

1573865459782.png
 
7,801
4,462
With one of those, the NSA should be able to plant a bug with a transmitter anywhere in the world and have the signal picked up in Fort Meade MD.
 
berkeman
Mentor
56,131
6,156
With one of those, the NSA should be able to plant a bug with a transmitter anywhere in the world and have the signal picked up in Fort Meade MD.
Moon Bounce?
 
Vanadium 50
Staff Emeritus
Science Advisor
Education Advisor
23,478
5,924
but what is up with the irregular spacing of the higher-frequency parabolic dishes?
You want all baselines and all angles covered. In principle you want a perfectly random distribution, but practicalities (like finite telescope size) preclude that. But it's close.
 
7,801
4,462
berkeman
Mentor
56,131
6,156
You want all baselines and all angles covered. In principle you want a perfectly random distribution, but practicalities (like finite telescope size) preclude that. But it's close.
Interesting. Does the VLA use random spacings along their regular tracks? And the phased array radar systems that I'm aware of use very regular spacing to make the math easier in combining the signals.

1573868398230.png
 
Vanadium 50
Staff Emeritus
Science Advisor
Education Advisor
23,478
5,924
phased array radar systems that I'm aware of use very regular spacing to make the math easier in combining the signals.
With radar you control the source and detector. With astronomy, you only control the detector.
 
berkeman
Mentor
56,131
6,156
With radar you control the source and detector. With astronomy, you only control the detector.
Very interesting, thanks. Is it like applying dithering to a noisy communication channel where the receiver has a low SNR? Is there an algorithm to optimally place the 2-d positions of the receiving parabolic dishes, or will any random 2-d placement be equivilent? I'd enjoy reading about the math behind this (from a communication channel perspective or otherwise).
 
berkeman
Mentor
56,131
6,156
And why doesn't the VLA use random track angles and spacings? Or does it use random spacings along their regular tracks for some observations? Thanks.

1573870695060.png
 
Vanadium 50
Staff Emeritus
Science Advisor
Education Advisor
23,478
5,924
I'm not a radio astronomer, but I believe VLA looks like it does because the distances and angles are somewhat tunable, which mitigates the problem, and it has many fewer telescopes so of you randomized it you would have a much smaller aperture. Making it the not-so-VLA.
 
berkeman
Mentor
56,131
6,156
Thanks V50. This is very interesting to me especially because I work with optimizing communication channels. It looks like they are doing something like dithering, but time invariant. Maybe spatial dithering or something. I'll keep looking for references.
 
berkeman
Mentor
56,131
6,156
Actually I should have paged @mfb since he should have pointers to references to the math behind the array spacings for this project... :smile:
 
Baluncore
Science Advisor
7,059
2,218
For radar only a pencil beam is needed. A regular array can be used to synthesise a raster scan beam simply by delaying rows and columns while summing the beams.

To generate an image of the source within a pencil beam requires a correlator and the ability to eliminate artefacts of the spatial array.

Early Radio Astronomy arrays of antennas were regular because all processing was narrow-band, analogue, on site, and in real time. Time delays between elements were programmed with coaxial delay lines and/or analogue phase shifters. By arranging an array with a N-S and/or E-W orientation it made it possible to scan the sky as the Earth rotated, or to track one point in the sky over a period of time.

The initial VLBI network employed what bigger dishes were available to get Earth dimension arrays for high resolution. The bandwidth was low as all data was digitised and recorded on magnetic tapes, one bit data, one track at the time, rewind the tape, record the next track, change tapes every few hours. Those tapes were sent airfreight to the correlator facility. Synchronising the data streams prior to correlation initially took most of the time because more accurate time references via GPS were not then available. A VLBI observation took 12 hours as the antenna locations rotated on the Earth, while each and every antenna tracked the same point in the sky. All spatial frequencies of the available antenna array would appear in the data record. That attenuated false correlation peaks while accumulating the hopefully true correlation image.

Waiting 12 hours means that half the observers will see the source rise, while the other half see it set. By using irregular arrays with low spatial self-correlation, but with a great total area, it is not necessary to wait for Earth rotation to spread the spatial array pattern to remove the artefacts of the array.

Small arrays, over maybe a few km span, at very short wavelengths, can be used locally for observation of absorption lines and chemical laser emissions. At longer wavelengths, those small regular arrays can combine their signals locally for use in a VLBI network.

In the last 20 years, faster A-D conversion with wider bandwidth have become available. GHz bandwidths can now be block down-converted, digitised and transmitted over fibre networks. That is what makes the SKA now possible.

Google Earth;
Australian “SKA Pathfinder” spreads out from here; -26.696971° 116.636942°
Australian “Murchison Wide Field Array” is close by at; -26.703033° 116.670424°
South Africa, the “SKA” is expanding out from here; -30.713472° 21.441458°
 
33,579
9,315
It is a bit like an inverse interference experiment. In a double-slit experiment you get intensity maxima wherever the phase difference is a multiple of the wavelength, which leads to a regular pattern on the screen. Similarly, two rows of radio telescopes would look at several parallel lines in the sky at the same time, and four telescopes arranged as square look at a grid-like pattern in the sky. The larger the distance between telescopes the better your resolution - but it means you get more grid points in the sky. If you are studying a bright source that is okay and you want to maximize resolution. If you study a dim source you need to get rid of all the other areas - you need more telescopes, and you want them closer together.

The VLA telescopes can move on rails to get the optimal position for a given observation. Putting all telescopes in three directions with railway lines looks like an efficient way to do this. I don't have maps for the resulting patterns, but I would expect that they are quite flexible.

SKA will look for very faint things with a huge number of telescopes. Moving them around would cost too much, scattering them in an irregular pattern is easier.
 
sophiecentaur
Science Advisor
Gold Member
23,883
4,090
Is there a reason for the apparent random spacing shown in the simulation/rendition video?
It's a so called Sparse Array which will give good resolution with less sensitivity than a fully populated array. The random spacing allows wide band operation because it suppresses bad sidelobes at awkward frequencies that would be susceptible with regular spacing but at the expense of slightly worse overall performance. Sidelobes are a nuisance because you need to look in several different directions to see the various peaks of reception before you can identify the highest peak which is the actual target direction.
A familiar version of a sparse array is the log periodic spaced antenna which has a good bandwidth and good directivity at the expense of gain. Ideal for TV reception but not necessarily in fringe areas. (It's not as simple as that because the TV version uses various lengths of element so the design does the 'thinking' for you).
 
Last edited:
sophiecentaur
Science Advisor
Gold Member
23,883
4,090
The bandwidth was low as all data was digitised and recorded on magnetic tapes, one bit data, one track at the time, rewind the tape, record the next track, change tapes every few hours.
Ahh. Those were the days my friend.
 
stefan r
Science Advisor
Gold Member
788
217
The inverse is also true. It is not possible to pick up airport radar with current capabilities less than 200 light years. After completing the km array airport radar at locations more than 200 light years will be too remote to detect.
 
33,579
9,315
It always depends on the distance. As extreme example, it is possible to detect airport radars light-milliseconds away. We don't expect artificial radio sources closer than at least 4 light years, of course.
 

Related Threads for: The Square Kilometre Array

  • Last Post
Replies
1
Views
1K
  • Last Post
Replies
7
Views
2K
  • Last Post
Replies
1
Views
2K
  • Last Post
Replies
8
Views
3K
  • Last Post
Replies
1
Views
2K
Replies
2
Views
9K
Replies
14
Views
2K
  • Last Post
Replies
6
Views
4K
Top