Using cheap satellite disks to make a radio astronomy baseline array

[SteveBz]

Hi Guys,

This a summary of an off topic discussion in a previous thread.  You can find the original here:

Thread so far:

SteveBz

... a small base line array based on satelite dishes from eBay might be fun.  There are a lot that go for 99p.  .......

...

But the really exciting thing would be if you could set up one, and I could set up one, and Robin could set up one and Jon could set up one ....  (eg 16 or 25 or 36 participants or more).

If we could encode the signals and combine them over the internet, imagine what we could have.  The mind boggles.

...

There are plenty of ways of synching information.  If we are synching at 21 cm, we don't need to send the carrier wave (ie 1.4GHz) we can just send regular packages showing the phase offset and amplitude and combine them across a peer-to-peer network.  Everyone would have access to everyone else's information.

...

robin_astro

Yep it is all about very accurate clocks (you need to sync the signals to a fraction of a wavelength so at say 1.4GHz you need  agreement between observers to perhaps a few tens of picoseconds) .  

SteveBz

Obviously you're right, thanks for pointing that out.  I wonder if there is a precise satellite signal that people could all pick up and share.

Then all you'd need is precise location.  I guess to maybe a tenth of a wavelength (1 or 2 centimetres).  The location would then be used to adjust the clock signal and calculate relative distances between dishes.

Regards

Steve

[Carl Reade]

Hi you like a challenge☺

Just a few thoughts for the mix, most folks will find this daunting as RA is not a common sport. How about putting together a standard kit or list of components and a user guide to get them started and take it from there.

Carl

 

[jazza]

I do this (interferometry) for a living (though less on the technical and more on the science side), and yes getting accurate clocks will likely be a major issue. 
Most VLBI  (Very Long Baseline Interferometry) stations use a hydrogen maser (atomic clock) as a frequency standard. At 1.4 GHz you can get away with using a Rubidium as your frequency standard but that's still a couple of hundred euros.

I guess nowadays at 1.4 GHz you wouldn't necessarily have to mix down much to be able to sample it but you probably still need a signal generator, some kind of a filter (not too hard) and then samplers etc.
I don't want to be discouraging but this is really really not easy.

On the other hand if you can run a cable directly between two dishes, all you really need is a mixer. Then you get a fringe pattern on the sky, like a double slit experiment you may have seen. So when the two signals interfere, where they line up you will see essentially double the signal (constructive interference) and where they don't you will get less signal averaging to zero (destructive interference).

It's pretty cool to see people doing amateur radio astronomy at any level
 

[SteveBz]

6 hours ago, jazza said:

I do this (interferometry) for a living (though less on the technical and more on the science side), and yes getting accurate clocks will likely be a major issue. 
Most VLBI  (Very Long Baseline Interferometry) stations use a hydrogen maser (atomic clock) as a frequency standard. At 1.4 GHz you can get away with using a Rubidium as your frequency standard but that's still a couple of hundred euros.

I guess nowadays at 1.4 GHz you wouldn't necessarily have to mix down much to be able to sample it but you probably still need a signal generator, some kind of a filter (not too hard) and then samplers etc.
I don't want to be discouraging but this is really really not easy.

On the other hand if you can run a cable directly between two dishes, all you really need is a mixer. Then you get a fringe pattern on the sky, like a double slit experiment you may have seen. So when the two signals interfere, where they line up you will see essentially double the signal (constructive interference) and where they don't you will get less signal averaging to zero (destructive interference).

It's pretty cool to see people doing amateur radio astronomy at any level
 

Very cool that you do this for a living (at Jodrell or where?). 

I think we'd certainly want to do a field test (or large garden) on the way to a distributed system, there are probably a number of steps:

1) A few individuals with dishes

2) A combined test in a large space with direct linking

3) .... more stuff towards a distributed array.

There was a bloke in Manchester selling 4 identical 1m dishes for 99p last month or the month before, sadly I live near Gatwick.  That would have been a nice test, no clocks or anything.

Regards

Steve

[johnfosteruk]

I would be very interested in this Steve. I've been reading some stuff on britastro and other sources for a little while but other commitments have taken my time recently. I may well get on the ebay this weekend though - I'll let you know.. 

I may even be able to convince the present Mrs Foster to let me out for a test day.. I'll be back on the thread again once I have some gear. 

By the way I know what you mean about Mike Brown's write up, very inspiring. 

[SteveBz]

12 hours ago, Carl Reade said:

How about putting together a standard kit or list of components and a user guide to get them started and take it from there.

First of all, to set your expectations, my total expenditure on radio astronomy to date is £24!  But that ignores the fact that I've 'borrowed' an EQ5 from my astronomy setup and that I already had a small multimeter with a USB connection, but actually I didn't use these at first as my first test was a total energy plot with a pencil and a piece of graph paper.  The kit looked like this:

Dish: go for the biggest you can find.  Mine is 1.3 m and has a basic and very rusty alt-az mount with a basic pulse modulated motor.  It was all very rusty and heavy and I had to collect it on the roof-rack of the car.  The guy just wanted rid of it.  The LNB still worked and it came with 25m of cable.  The cable was a combined coax for the signal and a 4 core telephone-type cable for the motor.  Cost £5.00 "collection only". 

Then you need a satellite finder, £4.99 but if you go for the £12.99 delux you get a whole lot of connectors and stuff too, it saves a lot of pain and wasted time.  The delux version is here:

http://www.ebay.co.uk/itm/361468620720?_trksid=p2055119.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT

A power supply for the finder.  I had two 9v rechargeable batteries plus connectors from Maplins which I strung in series to get 18v.  The slight problem is you leave your batteries charging overnight and then they run out in 45mins, so I used an old 12v power supply I had lying around.  It gives you a bit of mains hum, but nothing overpowering.  If you have a 12v car battery for astro use, that would be perfect: no noise, lots of oomph.

satelite dish - motorised (with plenty of cable, an LNB and a mount).	5.00
Analog Satellite Signal Finder Strength Meter LCD Display for Sat Dish DIRECTV	4.99
Coax connectors	4.17
AC 100V-240V Converter Adapter DC 36V 1A 36W Power Supply Charger DC 5.5mm New (see below)	6.89

My first measurements were un-driven, but when you become driven you'll need 36v for the satellite motor.  Make sure you get enough amps, I didn't and from memory I should have got 5A, but I got 1A.

If you don't get a mount with your dish, buy a dovetail bar and use your EQ mount.

If I've missed stuff off let me know and I'll update it.

Photos follow,

Regards

Steve.

Dish with "Permanent marker guidance system" - uses the shadow of the LNB to measure alignment.  If you have an offset parabolic dish (and most of them are), it will work differently.  Maybe you can mount something that casts a shadow on the opposite rim of the dish from the offset LNB.  Then it should work the same.

Antenna "control centre".

You can just read the value of the satellite finder, but to get a direct connection you can strip the back off and attach a multimeter like this:

Heres the reading on the multimeter:

Here's the output:

 

 

[robin_astro]

6 hours ago, SteveBz said:

I think we'd certainly want to do a field test (or large garden) on the way to a distributed system, there are probably a number of steps:

1) A few individuals with dishes

2) A combined test in a large space with direct linking

Hi Steve,

The problem with satellite TV dishes is that they down convert the signal to a lower frequency at the dish (The LNB contains an oscillator and a mixer). Because the oscillators are free running they will never stay in sync long enough to give interference fringes. You would need to lock them together or  feed them all  from a common master oscillator somehow rather than using the internal oscillators.  

This is easier to do at lower frequencies though as you can feed the signals direct to a central receiver without down conversion without too much loss in the cable and combine them there using a common receiver.  

If you dont have a suitable receiver and are looking to keep costs low, something like this might work, using two Yagi's instead of one. 

http://www.backyardastronomy.net/radio_telescope.html

 But note I am not convinced about the claim shown there to have detected Cas A using a single 10 element Yagi.  At that beamwidth (~25deg)  the background from the Milky Way would have dominated and I suspect what is shown there is actually a Milky Way transit. With 2 Yagis though you might be able to pick up fringes from point sources as in the link I posted on the other thread 

http://www.britastro.org/radio/RadioSources/extragalactic.html    Virgo A  (fig 11.11)

http://www.britastro.org/radio/RadioSources/supernova.html        Tau A  (fig 8.4)

Cheers

Robin

[Pippy]

Yes, to use a standard satellite LNB across your network of amateur receiving locations you MUST lock each and every LNB's local oscillator both in frequency and phase to each and every other LNB used in the network - this is going to be your first BIG task to overcome (the higher the RF frequency the harder it becomes).

After that, you MUST also do exactly the same (frequency and phase lock) for the rest of the receiving (any and all RF down conversions included) and digital sampling chain at each and every receiver location.

It's task you can only really begin to take on (hardware wise) once you've learnt about exactly what it is you're actually trying to do here.

You guys are talking about building a networked receiver system based on signal correlation .. Correlation being the key word.

[SteveBz]

1 hour ago, robin_astro said:

The problem with satellite TV dishes is that they down convert the signal to a lower frequency at the dish (The LNB contains an oscillator and a mixer)

You're right, but on my dish the LNB is easily detachable leaving a feedhorn and a dish antenna for very little money.  You're the expert, but I guess building a bandpass filter/amplifier for the signal would be the next step and that doesn't feel very expensive, do you think?

I read your link to the backyard astronomy website, which I think you may have mentioned before, but it's still very cool and it was certainly time for a revisit.  It mentions the CATV tuner ($24 all in), and maybe that would be fun to experiment with.  The nice thing about the dish is that is is steerable, and I've grown attached to it.  I'm not really ready to try the Yagi yet (although there is an old one on our roof, which I may cannibalise).  I like the fact that with very little effort the dish sits on my mount and it's driven.  If I polar align the mount, I can, within the resolution of the beam width, point it at a selected source.

Maybe the CATV tuner is in fact that bandpass filter/amplifier?  Is that what a tuner is?  Then I can get an output, run it down a coax and into the recorder. 

I don't think 2 dishes is enough, doesn't that just make it Young's slits in radio?  If I have, say 4 dishes equally spaced in a row, I'd get a strong central peak, like an Airy disk, which would be the object in focus.

You know those TV signal spliters which are, in fact, just an aluminium box with one hole on one side and several on the other, is that not just a mixer in reverse?  In fact like this:

Is that not my mixer with the ins and outs reversed?

The IN port goes to the recorder?

Robin, I may have got it all wrong, but do you think that's workable?

Regards

Steve

[Pippy]

This is what you guys are basically wanting to achieve (phased array) ..

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

As you can see, modern radar systems and modern networked RF astronomy now use the same system.   For current military, a flat radar antenna is simply an array of similar small antennas on a flat surface whose output is eventually correlated in order to produce a desired image / result. The correlation is normally currently done using FFT's (look into FFT processing once you have a working sytem).

Learning how to interprit real world (the universe et all) input using todays tech (which is still very limited) is a trying task that can become extremely fascinating once you realise the scale of what is required to do so - our science is no where ready to touch on self awareness and consiousness.

The process of learning all this in detail gives one a tiny insight into what the biological brain is actually doing with it's multiple inputs (which is somewhat beyond what we currently think we understand).

 

[Pippy]

16 minutes ago, SteveBz said:

Is that not my mixer with the ins and outs reversed?

The IN port goes to the recorder?

They're not mixers. They just feed an RF input to multiple destinations (IF receivers).

An RF mixer (or higher) up/down converts a signal from one input frequency to another (also requires proper filtering).

 

[SteveBz]

46 minutes ago, Pippy said:

Yes, to use a standard satellite LNB across your network of amateur receiving locations you MUST lock each and every LNB's local oscillator both in frequency and phase to each and every other LNB used in the network - this is going to be your first BIG task to overcome (the higher the RF frequency the harder it becomes).

Hi Pippy,

Thanks for your reply.  And it's a good point.

I had been thinking we would fix the frequency, at least for a single session, so I thought we might just need to determine the amplitude, phase and geographical location very accurately.  We could actually even assume that for any given moment the amplitude was constant for everyone, that way we would partially correct for the curvature of the Earth.

So then we would need phase and location with an agreed sampling frequency.

Regards

Steve

[SteveBz]

6 minutes ago, Pippy said:

This is what you guys are basically wanting to achieve (phased array) ..

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

Hi Pippy,

Very cool.  It has some Maths, which we've been a bit short on so far.  I'll have a look through it.

I'm off to a public astronomy session now, unless it's rained off, so I'll be back, either very soon, or more likely a bit later.

Thanks for your input,

Regards

Steve.

[Pippy]

29 minutes ago, SteveBz said:

I had been thinking we would fix the frequency, at least for a single session, so I thought we might just need to determine the amplitude, phase and geographical location very accurately.  We could actually even assume that for any given moment the amplitude was constant for everyone, that way we would partially correct for the curvature of the Earth.

You have to fix not just their frequency, but also their phase.

In this graph, the frequency is the same, but the phase between the two sine waves is different - 90 deg out of phase (ask about IQ processing at a later stage).

[SteveBz]

5 hours ago, Pippy said:

You have to fix not just their frequency, but also their phase.

In this graph, the frequency is the same, but the phase between the two sine waves is different - 90 deg out of phase (ask about IQ processing at a later stage).

Hi Pippy,

Are you sure about this? I thought it was the difference in phase that gives us our interference and therefore our resolution and pinpoint accuracy.  If we have two disks in the garden, say, or a hundred disks in the garden and you fix the phase, then all you'll be doing is summing their areas and maybe averaging out the noise, but not increasing focus, which is what a baseline array does.  If you have ten dishes equally spaced in a line 76 metres long, you'll get the resolution of Jodrell Bank, but not it's sensitivity, because the combined area is still only 10 square metres or so. In fact I'm not sure you would even get the sensitivity of 10 m^2 but maybe only 1 m^2, I'm not sure.

Robin, what do you think?

Regards,

Steve.

[Pippy]

1 hour ago, SteveBz said:

Are you sure about this? I thought it was the difference in phase that gives us our interference and therefore our resolution and pinpoint accuracy.  If we have two disks in the garden, say, or a hundred disks in the garden and you fix the phase, then all you'll be doing is summing their areas and maybe averaging out the noise, but not increasing focus, which is what a baseline array does.  If you have ten dishes equally spaced in a line 76 metres long, you'll get the resolution of Jodrell Bank, but not it's sensitivity, because the combined area is still only 10 square metres or so. In fact I'm not sure you would even get the sensitivity of 10 m^2 but maybe only 1 m^2, I'm not sure.

I'm just talking about how you need to first lock all your local oscillators in your dish network (in both frequency and phase). You don't need a 0 deg local oscillator phase difference, but you do have to maintain a fixed 'n' deg phase lock (phase differences can be compensated for so long as they are non-chanaging and known). Without first doing this you can't begin to combine your received outputs in whatever way you desire.

[robin_astro]

Steve,

Pippy is talking here about each dish having its own receiver (effectively the LNB) in which case the local oscillators would need to be phase locked so the relative phase between the signals from the different dishes  is maintained. (Look up superheterodyne receivers).  You need to do it this way with TV dishes as the microwave wavelength is too short to send down coax cables efficiently.  If we are talking lower frequencies using Yagis for example, you dont need to worry about this as you can send the unconverted signals down fixed lengths of cable and combine them directly at  a single receiver to measure the result.  Understanding the results though is still pretty complex. Even with multiple antennae you end up with a  complex fringe pattern rather than a single narrow beam as the target moves through the beam, dependent on the specific direction of the object relative to the individual antennae. This can then be interpreted in terms of the detailed structure of the target.   

Robin

[SteveBz]

Hello Guys,

Last night's public stargazing event (organised by the local council) was very popular we had about 200 members of the public and maybe twenty scopes of varying types and sizes, including my c8n.  Sadly the clouds stayed put and nobody saw anything except the scopes.  People were good humoured and we handed out brochures and I must have explained about EQ mounts and Newtonian telescopes about 20 or 30 times!  We met many people who wanted to sign up to the our local club, so I guess it was a success.  Maybe if we had been looking at the sky, we would not have talked as much or got so much interest in the club!

Ok, I didn't understand everything that you guys said:  I'm going to Google super-heterodyne.  I'm sorry if you feel frustrated at my lack of understanding, believe me: it frustrates me more!!  In the meantime I'm going to summarise the current situation as I see it.

It is this.  Today many, if not most people, have a satellite dish on their roof. Over the course of the next few years, people will change suppliers, move house or upgrade their dishes and there will be a constant supply of small, used dishes which will otherwise be heading for landfill or the scrapyard.  It would seem that that there is some sort of opportunity here.  We are also witnessing a convergence of technologies (Communications, Internet, TV, Satellite and video), that allows possibilities that would not have been possible in earlier years. I am proposing that there could be some sort of internet enabled peer-to-peer structure that allows people to sign up to and collaborate with others to view extraterrestrial radio or microwave sources in the form of an extended array.  I'm not sure that today we can do much more than look at our nearest neighbour the Sun, but standards are improving at an amazing rate and that may not be the case for long. 

So the main question is a two parter, 1) with these dishes, what is the easiest wavelength to detect and 2) what existing consumer technology would allow us to cheaply and easily get the attributes of that signal onto a PC where we can do something with it and share the result with others.  The attributes of any wave are frequency, amplitude and phase.  In an ideal world with no losses or delays we would capture all of these and share and recombine these in real time.  As we don't have a global microwave network, this remains outside our reach. However,  consumer video processing appears to me to show us a potential direction.  Video technology samples the attributes of visible light (frequency is reduced to R, G, B and we have amplitude) at 25 (PAL) or 30 (NTSC) frames (records) per second (fps) and it has three main benefits 1) it is cheap, 2) we can recreate the original image to an acceptable level of quality and 3)  we seem to be able to pass it across the internet with little problem.  If we could detect and encode selected attributes of our chosen Radio signal at, say, 30 records per second, then we have the basis of a collaborative interferometer.  Each computer will collect the output of the others and can interpolate between samples to recreate the original signal.

So let me ask you, what frequency would utilise the attributes of these dishes best and could then be easily processed with existing consumer technology?  We can always change it later, but agree what we are talking about here.  I imagined it was the 21 cm line, but Robin, you think a longer wavelength line might be better.

Let's put a marker in the sand for us to discuss.

Thanks for reading so far.

Regards,

Steve