Some of the world’s largest radio telescopes are not in fact as physically large as they claim to be, but instead are a group of telescopes spread over a wide area whose outputs are combined to produce a virtual telescope equal in size to the maximum distance between the constituents of the array. Can this be done on the cheap with an array of satellite dishes? It’s possible, but as [saveitforparts] found out when combining a set of Tailgater portable dishes, not simply by linking together the outputs from a bunch of LNBs.
The video below the break still makes for an interesting investigation and the Tailgater units are particularly neat. It prompted us to read up a little on real aperture synthesis, which requires some clever maths and phase measurement for each antenna. Given four somewhat more fancy LNBs with phase-locked local oscillators and an software-defined radio (SDR) for each one then he might be on to something.
If you’re curious about the cyberdeck in the video, you might like to read our coverage of it. And the Tailgater might be a bit small, but you can still make a useful radio telescope from satellite TV parts.
It might not be synthetic aperture and obviously phase alignment won’t work at all (without special handling) with LEOs and MEOs, one of the methods of improving fairly unreliable Ku and KA band links which are subject to severe rain fade is physically separated and parallel earth stations.
The two (or more) dishes are arbitrated for link fade and traffic bridged through the IP network.
The theory being that 2 physically separated terminals (2 to 3 miles) have different weather attenuation profiles and that more likely than not put one dish in the clear. Instead of an availability of 99.6%, predicted availability is in the range of 99.9 or better. I know it doesn’t sound like much difference, but it’s huge.
I understand they’re developing the same technique for space to earth based laser communications.
It doesn’t improve the aperture because link availability is local to the terminals and based on power, thermal noise, antenna aperture size, power spectral density, modcods, FEC, etc. But when rain fade on KA band can be 30+ dB, treating path fade as local dramatically increases the overall link availability.
Spatial diversity.
Succinct.
Correct.
The VLBI (Very Long Baseline Interferometry) using radio telescopes involve radio telescopes so far apart with some of them not connected with a good enough bandwidth, that the data is stored with VERY accurate time signals on hard disks and later combined.
Originally everything was stored on tape, that ran on till around 2010, disks replaced tape. It’s only in the last 10 or so years that the European VLBI Network (EVN) have been regularly been doing pure e-VLBI where data isn’t recorded but streamed to the correlator from all sites. However even now that’s only a small part of the overall setup (1 day a month-ish) and the longer sessions are all recorded. How this recorded data gets to the correlator depends on station connectivity – the preferred solution is online transfers but some stations do still ship disks.
With the very high observing bandwidths appearing (64 Gbps) shipping that over global networks isn’t likely to happen in the near future, so it’s back to disks.
How was this quote, something like “Never underestimate the bandwith of a truck full of tapes going down the highway”? And who said this?
It’s usually attributed to Andrew S. Tanenbaum and the original goes like this: “Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway”.
why this antena is circle not line?
The antenna is most likely a line, the part that is circular is a surface that is shaped to bounce all incoming signals towards a point in space where the very small antenna is put.
https://en.wikipedia.org/wiki/Parabolic_antenna
Is there a way to record the frequency “raw” at different locations, with some super exact time source, and then simply adding/averaging the signals (or better said: noise) and then get a better Signal to Noise Ratio?
As another reply said, the phase is also critical, because the interferometry is based on the phase difference. Some very tight tolerances.
Each of the four LNBs has its own downconverter(s). So, the scheme has at least four local oscillators (assuming each unit has only one local oscillator in its signal chain) which are running 𝙥𝙡𝙚𝙨𝙞𝙤chronously but not 𝙨𝙮𝙣chronously. I think what’s missing from this scheme is means to lock all local oscillators to a common reference, with means to introduce variable phase delays between units for best response. As it stands now, there is virtually zero chance the phase between the local oscillators remains constant, so the addition of the four signals is not coherent. This (plesiochronicity) may be what is causing the observed bands in the waterfall.
Exactly. Somehow, VLBI manages to encode all the phase data.
Didn’t expect someone to use the word “𝙥𝙡𝙚𝙨𝙞𝙤chronous”. Made my day.
I remember watching Charlie Sheen making one of these in the movie “The Arrival” from -96 and thinking, “It surely cannot be that easy”.
So I wrote an email to the people at the Very Large Array and got a two page answer within three hours, thankful, but thinking that things are probably slow at times for the people working at the VLA.
in 2011 Rocket City rednecks made their own array with dish network dishes. I looked all over the net for the software they used, but couldn’t find it. I think the one engineer wrote his own in matlab.
https://www.youtube.com/watch?v=2O_d-tgv28E