Variable sources

Rushton with Eyres and Beswick

Aim
The purpose of this tutorial is to investigate the effects of source variability on the image plane and produce a light curve from source visibilities. Your task will be to analyse MERLIN data of the X-ray binary Cygnus X-3 using AIPS.

Motivation
A fundamental assumption of aperture synthesis is that the true sky brightness distribution does not intrinsically change during an observation. Such an assumption is not true for highly variable sources such as XRBs; these sources exhibit large variations in flux density on the time-scale of a few hours over many days.

Data
The XRB Cygnus X-3 is known to vary between 0.1-1 Jy over a 12 hour period. Such rapid changes cannot be due to changes in the source structure and it can be assumed most emission originates from the central region. It is therefore desirable to study the light curve of the unresolved core so we can model the effects on the image plane.

Method

Download the uv fits file CYGX-3.FITS to a directory

Set an environmental variable pointing at the folder containing cygx-3.fits

in CSH

 > setenv DATA /path/to/folder    

or in BASH

  $ export DATA=/path/to/folder

Load the data into AIPS using FITLD

> task 'FITLD'

> inp 

> datain 'DATA:cygx-3.fits 

> go FITLD  

The data file has already been calibrated and contains only the source 'CYGX-3B'

 > pcat

Image the the uv data using 'IMAGR'

> task 'IMAGR'

> inp  

> getn #

> cellsize 0.01

> imsize 512

> niter 500

 > dotv 0

You can reorder the data using

> recat

Now select the image and view in the AIPS TV window

> getn #

> tvlod

As you can see the image has many 'ripples', this is due to variations in the source amplitude.

Check the baselines using VPLOT

> task 'VPLOT'

> inp

> getn #

> dotv 1

> bparm(2) -1

> go vplot

Point source variation will appear simultaneously on all baselines. Resolved structure will appear differently on each baselines depending on the separation of the two telescopes. Therefore if we average all the baselines together we will can make the most sensitive lightcurve of the varying curve. However, the task we use will vector average all the baselines (i.e. we must also consider the phase variance).

First we must find the centre of target source. Place a box around the source with

> TVWIN

and run

> go JMFIT

This will return peak of the RA and Dec deconvolved the the beam. Running

> imhead

will then return the pointing centre of the image.

Next the target source needs to be moved to phase centre of the image using UVFIX. The values AA and BB for shift are in arcseconds (i.e. remember to convert RA from HH:MM:SS) and can be dervided using:

RA' = RA + shift(AA)/cos(DEC)

DEC' = DEC + shift(BB)

> task 'UVFIX'  

> inp 

> getn # 

> uvfixprm(1) 1 

> shift = AA, BB 

> go uvfix

Check the source is at the phase centre of the image my re-running IMAGR

Run VPLOT again to check the phases on each baseline only vary by <40 degrees over the track.

Average together all the baselines using TBAVG

> task 'TBAVG'

> shift 0

> getn 1

> go TBAVG

Finally view the averaged baselines uv file using VPLOT

---

Optional – Improving the phase variation with phase-only self-cal

We need to check how many clean components are in the map

> invers 1

> go prtcc

We need to self-calibrated the target.

> task 'CALIB'

> inp

> getn 1

> get2n 3 % i.e. select the initial image

> niter XX % i.e. the ¾ of the clean components reported by prtcc

> solmode 'P!A'

> go 'CALIB' 

As before run IMAGR on the new uv file created by CALIB. The new image should produce a map with fewer 'ripples', but not completely error free.