fringe pattern show up after adding more extended configuration CARMA data

[Shuo Kong]

Dear all,

Sorry if this question is so naive.

I have a question about combining multiple configuration CARMA observations. So if I just invert D+E config data, the image look very "smooth", like the first figure shown below. But if I include C config data, lots of fringes show up (second figure). The C and D configuration data were observed the same amount of time, and I think the D-config data has much better SNR. So, my question is: Is this fringe effect simply caused by not enough sensitivity, or there is something wrong in the data?

Thanks a lot!

cheers,

Shuo

[Chat Hull]

Hi, Shuo--

No, the C-array data shouldn't look like that.  I wonder if you have a couple of bad baselines.  Notice the linear "washboard"-like ripples that go up and to the right (with a higher spatial frequency) and down and to the right (with a slightly lower spatial frequency).  That could be due to abnormally high amplitudes on a couple of baselines, or possibly on just one antenna (which of course is part of many baselines).

In short, I'd look at your UV data to see if there are any anomalously high points -- when I do this kind of testing, I usually look at amplitude vs. time of both the calibrator and the source, making sure not to do any time averaging just in case you had a high amplitude for a couple of seconds at the beginning of a scan.

Hope this helps,

Chat

--
You received this message because you are subscribed to the Google Groups "miriad" group.
To unsubscribe from this group and stop receiving emails from it, send an email to miriad+un...@googlegroups.com.
To post to this group, send email to mir...@googlegroups.com.
Visit this group at http://groups.google.com/group/miriad.
For more options, visit https://groups.google.com/d/optout.

--

Chat Hull, Ph.D.

Jansky Fellow
Harvard-Smithsonian Center for Astrophysics

National Radio Astronomy Observatory

LinkedIn: linkedin.com/in/chathull

Web: chathull.com

[Peter Teuben]

Shuo,
 
    one can clearly see an underlying map in the D+E in the C, so the
basic data is ok, but from the pattern I can see there must be some
bad data not flagged yet in the C data. Could also be something with
a calibrator of course, messing up the real data. Or a combination.

One way I like to see patterns in time or frequency (or baseline) is
to use the uvimage program, and use progams such as ds9 to
visualize the resulting image cube. Perhaps you can see odd patterns
in there. or time range or certain antennas that are bad.
You can use the closure program to see if there are phase errors
in the calibrator. And of course you can inspect the ph/am plots with
uvplt/uvplot,and spectra with uvspec(t).

peter

[Shuo Kong]

Thank you all for the quick reply! It's weird because I thought I flagged out suspicious data, but I'll check it again. Your information are helpful!

--
Sent from Gmail Mobile

[Lauren McKeown]

Hi Shuo,

Oh wow, my first radio astronomy question! 

Unfortunately I'm out of radio astronomy a while (moved onto planetary science) but hopefully I can help you with this as I had similar problems with my observations of red giants last year. First of all - what source are you observing? Is it extended or point source - usually the longer baseline arrays are used for point sources, as the angular resolution is better. 

Looks like you have a lot of noise in the latter map...did you observe at 1 mm or 3 mm? I found it extremely difficult to get clean maps for my sources observed at 1 mm as the data was affected by sky rms and water vapour; this was especially the case for a FAINT giant star which I observed at 1 mm. 

Lastly - how much time did you leave between observations? I'm presuming between E, D and C config this was a couple of weeks? Might be good to check the weather reports at Cedar Flat for your observations just to be sure - especially at 1 mm. If it is the case that the rms and tau 230 exceed CARMA's recommended values for C config, I'd concentrate on your E and D config data. There may be some handy MIRIAD tricks to clean up that C config map but I'm familiar with CASA I'm afraid and only used MIRIAD at the CARMA Summer School.

Hope I've helped somehow!

Lauren.

--

[Lauren McKeown]

I only did a few months last year of this stuff though by the way and am starting my first year of a Ph.D in an entirely different research area...so please take anything I've said with a pinch of salt! The experts in this group will know much better :) I hope you solve the problem and wish you all the best with your work! 

[Jackie Villadsen]

Hi Shuo,

Also, are you using uniform or natural weighting when you use invert to produce the dirty image?  If you are using uniform weighting, one or two baselines in a region where there is not much uv coverage could be producing the fringes.  (I haven't used the CARMA Large Array so not sure if this is a concern here.)  See for example the first two images in the "Results (dirty images)" section of this web page:

http://www.ast.uct.ac.za/~ims/aufgaben/meerkat_PB.html (these are simulated interferometric observations - different array and different software, but it illustrates the potential effect of uniform vs natural weighting in the case of sparse uv coverage)

If you are using uniform weighting (running invert with robust <~ -2), try using natural weighting (robust ~ 2) instead and see if that helps.  If you do not specify the value of robust when you run invert, the default is uniform weighting (robust = -infinity).

As mentioned above, it could be a problem with your calibration.  Have you tried imaging your phase calibrator for your C array data?  This can help you narrow down where you should be looking for bad data.  If the phase calibrator image looks clean when you image it using the same settings as you used to image your target field, then look for bad data in your target field; if the phase calibrator image has ripples in it like those in the target image, then look for bad calibrator data that is causing a calibration error.

Jackie

[Shuo Kong]

Dear all,

First of all, thank you for helping me! 

I find out the problem, it's a bad antenna. Now the map looks better. However, I don't see any weird things in all kinds of plots. "uvimage" is a great tool (thanks Peter!), the baseline-channel images show no weird things, except that some of the visibilities are zero (which I think they are flagged?). 

Chat, I plot amp/phase vs. time but don't see weird things (actually I'm using the calibration script from John Carpenter from CARMA Summer School, and it plots lots of figures). 

Lauren, I'm using 3mm, and the tau230 looks fine.

Jackie, I used "sup=0" so I think it's natural weighting?

Anyway, what I ended up doing was flagging antenna one by one, trying to find maybe one or two that have the problem, and it worked! I had 2 C-config observations, both look good if I flag ant(3). This method is kind of dumb, but it worked. However, I don't recall anything weird about ant(3) during calibration. But maybe I don't plot the right thing or I don't recognize the bad data.

cheers,

Shuo