Jump to content

Stargazers Lounge Uses Cookies

Like most websites, SGL uses cookies in order to deliver a secure, personalised service, to provide social media functions and to analyse our traffic. Continued use of SGL indicates your acceptance of our cookie policy.

sgl_imaging_challenge_banner_solar_25_winners.thumb.jpg.fe4e711c64054f3c9486c752d0bcd6f2.jpg

RobB

Radio Intensity Interferometry at low-ish cost

Recommended Posts

Been wanting to expand my mind into Radio Astronomy and Astro Physics for a while.  Attended a IET lecture recently that nudged me to do something about it.   So what is required to build a radio telescope that can do a bit more than tell you that we have a sun, moon, Jupiter and a tantalising glimpse of possible spiral arms?  i.e. How to get higher resolution on an amateur budget (£2k-£3k).

Have started looking into what is required for Radio Interferometry.  Scouring the internet has turned up quite a lot of info. However, some of the more interesting amateur works seem to have gone dormant.  Started this thread to try and pull together the information sources, if only for my own benefit, but preferably others as well.

Seems to be two types of interferometry open to amateurs:

  1. Locked Local Oscillators (LO) and matching phasing harness for the antennas
    • requires antennas to be in close proximity
  2. Intensity Interferometry - No LO issues, useful over longer baselines.  Fair amount of Maths.
    • may require a fair amount of data traffic and processing power.
    • could be expanded to other amateurs (see project ALLBIN)

My preference would be to go down the "Intensity Interferometry" route and build on the work by others where I can dig it up.

Internet Sources (to be added to):

European Radio Astronomy Club -

  • Basics of Interferometry  - no LO synchronization necessary (slide 23). 
  • Based around a project called ALLBIN.
  • Link to membership or message board no longer valid.  All gone very quiet.

David Lonard

David Wilner - Harvard-Smithsonian Center for Astrophysics - Very good YouTube videos

Edinburgh University

University of Amsterdam

BOOKS

  • Interferometry and Synthesis in Radio Astronomy - Thompson - Third Edition
  • Tools of Radio Astronomy - T.L Wilson - Sixth Edition

Misc

  • Like 1

Share this post


Link to post
Share on other sites

Some great links there Rob.  It prompted me to investigate Intensity Interferometry which is a technique I had not come across before, either in radio or optical astronomy. I did not realise that the stochastic variations in photon counts (and low frequency variations in radio signal intensity) are correlated between different telescopes. 

Robin

Edited by robin_astro

Share this post


Link to post
Share on other sites

Another item to note.  No synchronisation of Local Oscillators required as long as you are direct sampling. I.e no LO required.  So 10GHz might be difficult between sites if required.  

Time stamping of the IQ feeds needs thinking about. I say IQ because you want to time stamp before a non-real-time operating system gets involved.

I have both the books I listed above.   NOT light reading.  Lol.

David Wilmer’s YouTube videos are particularly good.  One thing I do think is critical is the calibration of the gain of the system which needs to be carried periodically during a scan to remove temperature issues.  Also different size dishes (different beam width) would need adjusting for.

Share this post


Link to post
Share on other sites

From a historical perspective I have just started reading Hanbury Brown's book from 1974 "The Intensity Interferometer" which tells the history of his work developing the technique  initially in the radio and then the optical, downloadable from here for example

https://electrooptical.net/static/oldsite/hanbury/The_Intensity_Interferometer-Hanbury_Brown.pdf

The initial scepticism among some physicists at the time and their attempts to disprove it could work are interesting. Sadly in the optical at least, compared with phase/amplitude interferometry the correlation in the signal is very low so you need pairs of large telescopes to measure even the brightest stars and the technique has pretty much been forgotten.  Is the situation different  at radio wavelengths?

Robin

Edited by robin_astro
  • Like 1

Share this post


Link to post
Share on other sites
43 minutes ago, RobB said:

Another item to note.  No synchronisation of Local Oscillators required as long as you are direct sampling. I.e no LO required.  So 10GHz might be difficult between sites if required.  

As I understand it (and I am willing to accept I probably dont!), with Intensity Interferometry you are just looking for  correlations in  the demodulated signal so does how you get there eg phase locking any down converters, matter? 

Share this post


Link to post
Share on other sites

That might be my confusion.   Been investigating a fair few different VLBI solutions over the last few days. 

The solution that David Lonard was looking at included a step whereby the feeds were synchronised (don't have the full details yet) in his python FX Correlator code prior to running FFTs in the GPU.  This was with a system where both RF ends of the baseline were coming into a single dual channel LimeSDR.

He also started looking at GPS Disciplined Clocks.  See VLBI Oscillators.  My assumption is that, even with a perfectly synchronised timebase, a separate Local Oscillator for down converting would introduce an unknown phase shift at the remote site unless the LO is phase aligned with the synchronised timebase.  I'm getting a little (huge understatement!) out of my comfort zone here but that's what learning is all about :)

Share this post


Link to post
Share on other sites

Interesting stuff. I was thinking of building a Phased array with three double-biquad Antennas for Pulsar observing. This guy detected a Pulsar with reasonable S/N ratio for very little cost: http://neutronstar.joataman.net/sites/iw5bhy_barga_3/index.html

 

If I add up three of these antennas (one right next to the other), I should get a higher S/N ratio. But the beamwidth will be narrower, which is not a good thing when I'm integrating over a long time and the antennas are stationary... (Still looking for a way to steer my 1.5m dish antenna)

Share this post


Link to post
Share on other sites

I had a quick chat with David Lonard and he has used a GPS disciplined clock for 50MHz but didn't go very far, although did prove it can be done for an amateur.  At 1420MHz the issues will be magnified.  He's now concentrating on locally separated receivers (i.e. RF comes back to one point with one central clock).

I'm not convinced the pure Intensity Interferometry (no clock synchronisation) is the way to go (ready and willing to be corrected on this), so solving the clock on a budget is the gateway to VLBI for amateurs.  Interestingly he did point me towards XTRX Octopack which includes a GPS and can timestamp the data stream.  GPS needs to be married to another technology as I assume the phase noise and jitter will be too much (p.s. all new to me).  Rubidium maybe (I have 3, all of which need a bit of tlc).

 

Share this post


Link to post
Share on other sites

Can someone explain the role of the GPS in the receiver system? Are these GPS disciplined clock devices used to accurately measure the time of the incident waves (=phase of the waves)? If so, how do they achieve this? I'm not 100% sure, but I think professional VLBI interferometers use atomic clocks to measure time accurately.

The PrimaLuceLab H142-One receiver has also got a GPS device, whose role I don't fully understand.

Share this post


Link to post
Share on other sites

Coto - I think the H142-One receiver has a GPS device to stop any frequency drift from the internal local oscillator.   i.e. when you looking at 1420MHz it really is 1420MHz your looking at and not drifted to 1420.10MHz once it's warmed up.  The clock defines the sampling rate of the SDR.  The 'disciplined' part of 'gps disciplined' removes long term drift effects of temperature etc on the local oscillator, while the local oscillator handles the short term stability.  And the discipline part needs to done in a way that doesn't cause phase noise (i.e. correcting frequency in sharp jumps).

For radio interferometry it will do the same job, i.e. the frequency drift is reduced to almost zero.  However, as you guessed, it does another role and therefore needs to be a lot more accurate and reproducible.  Very accurate timestamp so when we recombine the signals at the computer the signal coming from the same direction we're interested in are constructively summed (need to delay one data stream in software to match the other due to path difference) whereas signals from other directions are not.  I'm still trying to work out the accuracy required (% of wavelength time.. for 21cm this is 0.7ns) to give a small enough phase error for an acceptable result.  Also, is the timestamp from GPS accurate enough (the 1pps output, is it at the same time at both sites)?

Also need to accurately plot where the two aerials are in relation to each other to a fraction of a wavelength.  Not so easy for us amateurs if not within a few hundred meters.  Military GPS receiver or averaging civilian gps position over an extended time?  Anybody a surveyor here who can tell me it's easy?

GPS is of course derived from atomic clocks (in each satellite and at the ground base station).  

Share this post


Link to post
Share on other sites
On 14/01/2019 at 21:29, robin_astro said:

From a historical perspective I have just started reading Hanbury Brown's book from 1974 "The Intensity Interferometer" which tells the history of his work developing the technique

Ah, straight from the horse’s mouth, so to speak... not forgetting the contribution of Twiss, of course.

Share this post


Link to post
Share on other sites

How do these GPS devices stop/minimize frequency drifts? When you say GPS don’t you mean those location-detecting devices? I don’t understand how such device can help with frequency stability or even be used as a very accurate clock..?

Share this post


Link to post
Share on other sites

It's a bit long but have a look at this YouTube video - GPS Disciplined Crystal Oscillator.  The way GPS works is based on very accurate clocks so you can 'measure' the distance from yourself to a number of satellites. Trigonometry does the rest.  To do this the GPS satellites transmit a very accurate timebase/clock.  As a side effect the receiver (the one with you) can provide a very accurate 1 second 'tick', the rubidium or crystal (or even cesium) oscillator fills in between each tick.  So the oscillator gets a very small correction every second removing thermal drift and aging errors.

Edited by RobB

Share this post


Link to post
Share on other sites

Interesting.. so how do GPS devices compare with local atomic clocks they use in professional interferometers? Are they slightly less accurate?

Share this post


Link to post
Share on other sites
3 hours ago, RobB said:

Also need to accurately plot where the two aerials are in relation to each other to a fraction of a wavelength.  Not so easy for us amateurs if not within a few hundred meters.  Military GPS receiver or averaging civilian gps position over an extended time?  Anybody a surveyor here who can tell me it's easy?

Would something like this help?  I found it a while ago when I was trying to find a way of getting polar alignment from GPS...

https://www.instructables.com/id/Sub-Centimeter-GPS-With-RTKLIB/

I didn't do anything with it - I'd need a few months to understand it all neer mind do it!

Michael

Share this post


Link to post
Share on other sites

The standard for VLBI is Hydrogen Maser.  GPS is then used to timestamp the data streams.   Rubidium can be used below 1GHz as the errors due to ionosphere path delays are greater than any timing errors due to the Rubidium, and can also be used as a backup.  Not sure about a GPS disciplined Rubidium oscillator.

Hydrogen Maser drift: < 10 to the power -15 Hz per day

Rubidium drift: 10 to the power -13 Hz per day  [how can you superscript here?]

And if your beam happens to encompass a pulsar you can use that...it will be a lot more accurate!  Unfortunately our dishes won't be big enough :(

p.s.  Trying not to sound like a know-it-all....I've just read this in section 9.5.3 of "Interferometry and Synthesis in Radio Astronomy:)

Edited by RobB

Share this post


Link to post
Share on other sites

Hi, just as a crude example this is what I had to do to keep a superheterodyne circuit board frequency stable for a magnetometer.

The circuit board was inside a thermos flask placed in the cooler box with a heatmat and sealed then controlled to within a degree of set temperature with a digital thermostat. All that for a 2 X 2 inch board?

 

IMG_20170326_232025.jpg

Share this post


Link to post
Share on other sites

Some of the GPS Disciplined Oven controlled crystal oscillators I've seen so far need the unit turned on for hours, if not a day, to fully stabilise.

Share this post


Link to post
Share on other sites
32 minutes ago, RobB said:

Some of the GPS Disciplined Oven controlled crystal oscillators I've seen so far need the unit turned on for hours, if not a day, to fully stabilise.

Yes my project required nearly 48 hrs but was then on 24/7. A time consuming businesses.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.