Polarisation Calibration - CABB data

Got a calibration problem? Discuss it here.

Moderator: Mark.Wieringa

Post Reply
julie
Posts: 12
Joined: Tue Mar 22, 2011 2:53 pm

Polarisation Calibration - CABB data

Post by julie »

I have begun to processing the ATLAS CABB data, observations completed in 2009 - 2010. I am looking at attacking the polarisation calibration and I have hit a bump. Our observations span 1 - 2 GHz centred at 1.4 GHz. Since leakage is frequency dependent, it has been suggested to me to split the band up into smaller chunks and determine the leakage for each frequency chunk.

Instead of creating multiple uv data-sets for each frequency chunk, can I use the "line" parameter in gpcal and select the channels for which I want to solve for the leakage? For example, work on channels 1 to 100, then on 100 to 200, etc.

gpcal vis=mydata refant interval options=xyvary,qusolve line=channels 1 to 100
gpcal vis=mydata refant interval options=xyvary,qusolve line=channels 100 to 200

... etc.

Does the calibration table get overwritten each time I run gpcal, making the calibration for channels 1 - 100 to be replaced by channels 100 - 200?

cheers,
Julie
ste616
Site Admin
Posts: 220
Joined: Thu Feb 04, 2010 3:27 pm
Location: Paul Wild Observatory Narrabri NSW

Re: Polarisation Calibration - CABB data

Post by ste616 »

Hi Julie,

You've asked a very timely question. Miriad datasets can only hold one calibration solution each, and each calibration solution can only have one leakage table, and that table is frequency independent. Therefore gpcal will be overwriting this leakage table each time you run it, and forgetting about the previous solution from the other set of channels (sort of anyway, it will use the previous leakages as an initial guess). I am afraid that for the moment at least (we are currently looking into expanding Miriad to have frequency-dependent leakages) you will need to split the band into separate sets to investigate polarisation properly.
cheers
Jamie Stevens
ATCA Senior System Scientist
julie
Posts: 12
Joined: Tue Mar 22, 2011 2:53 pm

Re: Polarisation Calibration - CABB data

Post by julie »

Thank you for the quick reply. I'll continue down the track of splitting up the data into frequency chunks. Any idea when miriad will be able to handle the multiple calibration solutions required?
julie
Posts: 12
Joined: Tue Mar 22, 2011 2:53 pm

Re: Polarisation Calibration - CABB data

Post by julie »

I see that gpcal was upgraded to be able to solve for the leakages without having to split up the band.

With nfbin>1, GPCAL writes additional tables 'gainsf' and
'leakagef' with solutions for frequency binned data.
Not many other tasks currently read these tables, but
they will be applied in preference to the single frequency
versions if present.

My question is does gpboot, mfboot, and gpcopy read these new leakage tables? Or should we continue to split up the band to apply the polarization calibration properly?

Cheers,
Julie
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Polarisation Calibration - CABB data

Post by Mark.Wieringa »

Hi Julie,


> My question is does gpboot, mfboot, and gpcopy read these new leakage tables?

Yes all these tasks read the new tables and will update/copy them.

Some of the less often used programs like gpaver, gperror, gpnorm have not been updated so they will just look at the continuum solutions.
Whenever the calibration is applied the new tables are read, and a message to that effect is printed.

If all is working well the new scheme should be better than 'divide and conquer' since the binned solutions are now interpolated to each channel. This should avoid steps in gain/phase or leakages at the bin edges. It also solves for delay by interpolating the binned phases.

Cheers,

Mark
julie
Posts: 12
Joined: Tue Mar 22, 2011 2:53 pm

Re: Polarisation Calibration - CABB data

Post by julie »

The ATCA CABB Polarization memo has been placed on the web at

http://www.atnf.csiro.au/observers/memos/

The memo is number AT39.9_129 and can be found near the bottom of the page. This memo discusses splitting up the band into smaller frequency chunks to calibrate and what one can expect for on-axis instrumental polarization. However, it should be noted that GPCAL will now split up the band for you using the nbins parameter.

Cheers,
Julie
Post Reply