C Band Interferometer

[Brett Dawson]

Hi David,

I've just stumbled across this group and have learnt a lot in short time!

I have been working on a C Band Interferometer (2 x 1.7m dishes - prime focus with chap feeds, 10m baseline, C Band LNA's) and have come up against many of the issues you've noted in this group.  I have been successful in producing fringes for a number of sources.  I've used phase switching and analogue correlation with an RF mixer (thanks Ken Tapping), but would like to use an SDR and move away from these analogue techniques - hence my interest in your progress with LimeSDR.  I have an RSP-Duo which I'd like to use, else may look at following your lead.

One of the problems I'm facing is gain instabilities in LNA's.  I have 2 x EF Data LNA's, one of which is quite stable with temp, but the other is not.  I'm looking for alternatives and I noticed you've been using modified Norsat LNB's.  These appear to have the waveguide to coax transition probe offset to one side.  Can you explain why this is offset and how it changes the performance of the LNB?

Thanks,

Brett

[David Lonard]

Hello Brett,

The Norsat feeds are designed with the probe to the side and with a tapering/narrowing in the E plane to increase bandwidth. My understanding is that it sets up a dual resonance that allows it to achieve greater bandwidth than a simple rectangular or cylindrical feed horn. This likely improves their temperature stability since the feed dimensions are no longer so critical. I only modified my Norsat LNBs in that I removed their 10 MHz reference so that they can accept an external 10 MHz clock.

 With 2.1 meter dishes, I can detect strong sources (Orion A, Taurus A and Sgr A), but I can't seem to get much more than that.

The upper C band is also very free of RFI, so an analog interferometer might be more feasible with this LNB if you could utilize the entire passband of a Norsat upper C band feed (~350 MHz). I'm playing around with an AD8304 phase comparator that logs to a LabJack that might do the job. 

David

[Brett Dawson]

Hi David,

thanks for the explanation of the offset probe arrangement.  The NORSAT LNB's look good, though freight to Australia is expensive. 

Why do you think only strong sources are visible? Is your Tsys as expected?

I am using an Ithaco lock-in amplifier then to a Labjack with Radio Sky-Pipe.  I'm interested to hear how the phase comparator goes and what sensitivity is achievable.  Will that give a linear output with phase?

Cheers,

Brett

[David Lonard]

Brett,

I've got a Tsys that is expected and I can generally tell that the system is very sensitive when I can detect fringes sun fringes leaking in from the dishes side lobes. I suspect that nearby objects like my house and some trees could be causing problems, particularly if they are coherent to the interferometer. I've been thinking about building a shroud for each of my dishes to reduce this potential problem. C-BASS uses shrouds, that while they increase Tsys, it blocks interfering signals that might sneak in through side lobes. I am also going to concentrate on Taurus A going forward since it is near the zenith at my location. 

From what I can tell, the Analog Devices phase comparator should be pretty robust and linear. It also provides phase and amplitude.

David

[Brett Dawson]

David,

I am looking for catalogues of bright 4GHz sources - I've used https://vizier.cds.unistra.fr/viz-bin/VizieR-2 to track down a HII survey down by Paladini et al, but am looking for others - do you have any ideas - what are you using?  As I'm in southern Australia, I've looked for work from Parkes etc but they seem to not record the brighter sources.

I have ordered some Norsat LNB's to get away from C Band sat interference.  Taking a while to get here so no progress yet.

Yes I am seeing side-lobe issues as well but only with strong sources - for example the sun and geostationary satellites.  With antennas near zenith, I often see what I assume is the rising sun in side lobes.  I see an effect where a fringe pattern develops, but the frequency of the pattern changes (increases with time).  I assume this happens because the foreshortened baseline length increases with time, and eventually the pattern disappears into the noise.

I've attached an plot of a recent result of M17. 

Thanks,

Brett

[David Lonard]

Brett,

I've mostly focused on sources that are listed here that were observed at 10 GHz.

http://www.ham-radio.com/sbms/presentations/SARA_projects_for_radio_amateurs/Interferometer_V2.pdf

Also, I looked at this effort to pick out strong C band sources:

https://www.radioastronomysupplies.com/uploads/9/2/4/8/92482330/radio_astronomy_experiments_at_4_ghz.pdf

I've got good fringes from M17 myself that look a lot what you are seeing in your plot. I think that this largely shows how challenging it is to detect fainter sources. M17 and the other handful of sources are very strong, but there is a rapid drop off in flux from virtually everything listed in the Parkes surveys or in other catalogs.  If these other sources are 0.1 to 1% the flux of M17, they are going to be lost in the background noise that you see on either side of the M17 fringes. One thought I have is to use a regulated heater to keep the LNBs at a fixed temperature to improve stability. 

Are you using a SDR and digital correlation?  I also notice an effect at sunrise and sunset that could be sidelobe related, or possibly temperature dependent effects on my coax and/or LNB.

David

[Brett Dawson]

David,

Yes the weaker sources are difficult - I've found that in some cases 100 Jy sources are detectable, but in others the same flux density is not detectable. Unsure if this is a systemic issue (eg pointing error), problem with catalogue data, Radio Eyes or something else.  I do know I need to improve Tsys, and so hope that the new Norsat LNB's will be better than the EF Data LNA's I am currently using (3.6-4.2 GHz band) 

I have observed about 5 or so HII regions that are bright and fairly easily detected at 4 GHz.  I think there are more, and I am putting together short list as I observe them.

I've attached an extract from Paladini (2003) for HII regions - sources are > 100 Jy (I think about 40 sources).  If you have Radio Eyes, you can load this as a new source file. 

I'm using an analogue phase switched interferometer.  I definitely want to start down the SDR track.  I see there's a long wait time on LIMESDR at the moment, so maybe I'll look for something else.

I'm struggling to understand how I can accurately measure the phase difference between my E and W antenna cable/down conversion systems.  That, I think, is my next challenge...

Cheers

Brett

[David Lonard]

Brett,

Thanks for the file, I'm pretty sure the 100+ Jy sources will be detectable. I get nice fringes from Orion A that lies on the Clarke Belt at my location that makes pointing easier. I agree that pointing errors are a hard issue to handle.

With the LimeSDR or XTRX I can see delay issues in the left and right receivers due to a sloping phase angle across my phase spectrograms. I posted an article about this before. With a analog setup this will be harder to measure. To get my coax cables just about right, I use a time domain reflectometer feature from the Mini1300 VNA that I normally use for antenna tuning for amateur radio.

I'm planning on focusing on the Crab Nebula in the future since it is near the zenith for me plus some thermal regulation. 

David

[Brett Dawson]

G'day David,

just thought I'd check in with you and let you know I've changed to the Norsat LNB's and have found them to be a significant improvement over the previous LNA's and separate d/c system I was using.  They are more (gain) stable, and likely have a better noise temp (not yet measured though).  I have also moved from 10MHz to 120 MHz bandwidth, centered on 4.8 GHz.

Attached is a recent obs of Orion.

Cheers,

Brett

[David Lonard]

Brett,

That does look really nice. I think it would be interesting to stack the Orion A observations over a few days to see if the fringes strictly follow a sidereal day. 

Best regards,

David

[Brett Dawson]

David,

yes I agree - I'll put something together.

Cheers,

Brett

[Brett Dawson]

David,

here are data from 2 days - 6/11/21 and 7/11/21.  The time UT.  The plots are offset by 4 minutes as expected. Original data attached as well.

These plots show truncated negative going peaks because the OP amps were being pushed a little too hard...

Cheers,

Brett

[David Lonard]

Brett,

It looks promising to me. The noise on either side of the Orion A fringes appears to be random and along the same average level, such that with averaging it will hopefully cancel out which could reveal weaker sources. 

David