Manu3C147_log

Through this log many of the new features of the new version 1.1.1 of Meqtrees will be pointed out.

Working on birch. Load 3C147 WRST data from http://www.astron.nl/meqwiki/OlegSmirnov/3C147CalibrationTutorial

Copy the preflagged 3C147_spw0.MS and the model 3C147.MDL in a working directory 3C147. In the same directory copy the TDL script calico-view-ms.py found in /usr/lib/meqtrees/Cattery/Calico.

MeqTrees must be run in a directory you can write to, such as your home dir or in our case in the working dir 3C147. Run the command: $meqbrowser &

FIRST LOOK: The aim of this step is to become familiar with the browser and with the data, to see how they look and to make a first dirty image without applying any calibration. Use the option TDL to download the TDL script calico-view-ms.py. Remember to tick the option " Change to directory of TDL script when loading". Set the TDL Compile time options as showed in the attached screenshot 1.1. Consider only the diagonal terms XX and YY. In the field "Interferometers to use" select the option S83, this excludes the antenna 6, redundant short baselines and some well known noisy baselines of WRST. Compile the TDL script.

Set the TDLJobs & Runtime Options as for the attached screenshot 1.2. With respect to the previous version here the field Read flags from MS the parent box and the child boxes should be ticked if you want to take into account of the flags. Even if not mentioned the algorithm used for the imaging is lwimager. Use again the S83 option in the field Interferometer to use. Check the image and plots using Bookmarks clicking ViewMS in the TDL Exec.

CALIBRATION OPTION I: In the next steps we will calibrate the DATA comparing them to a MODEL. Load the script calico-wsrt.py. Before to compile it read carefully the detailed description of the parameters which appear labeled with "i" in TDL Compile-time Options.

The differences between version 1.0 and 1.1 are the the following: "Antenna subset"--->"Interferometer to use" "What do we want to do"--->"Processing options" "Calibrate on: visibilities" ---> "Calibrate on" complex visibilities" "Subtract sky model and generate residuals"--->" Output visibilities" "Correct the data residuals" ---> "Output visibilities" "Show LSM GUI"--->missing "Use interferometer gains"--->"use multiplicative IFR errors" & "Use additive IFR errors"

In order to get a subtracted calibrated image, therefore to be able to see multiple sources in the field, we need to select: "correct residuals" in the field "Output visibilities".

On top of the parameters described there are few new in the tree Sky model: "Use 'Calico.OMS.central_point_source' module "Use 'Siamese.OMS.fitsimage_sky' module "Use 'Siamese.OMS.gridded_sky' module

Complete the TDL compile options as screenshot 1.6 and 1.7 and compile.

CALIBRATION OPTION II: TDL Jobs & Runtime Options should be compiled as screenshot 1.8. Not different options with respect to version 1.0. Screenshots 1.3 and 1.4 are the G_inspector_data and the correct_inspector plots. Screenshot 1.5 shows the dirty image of the calibrated and subtracted data.

SKIP THE RESIDUAL FLAGS

SOLVING FOR CLOSURE ERRORS: In the TDL-time Options add: Use Interferometer gains ---> Use multiplicative IFR errors (closure errors) Calibrate on: using interferometers: all The image is shown in screenshot 1.9.

Questions related to the three steps above: Q) How can we see in a user friendly way our model? Like coordinates, positions etc? How to plot it? A) Use the new tool Tiger (ask Oleg for the README and details)

Q) Where we see the option which tells you how many sources we are subtracting? A) In this tutorial we subtract all the sources included in the model, but there is a parameter in TDL Compile-time Options in the block Sky model which mentions Use subset of LSM sources.

Q) Is it possible to monitor the intermediate steps during the calibration? E.g Image after calibration only, image after calibration and subtraction? Do you have to calibrate every time? A) No, you do not need to calibrate every time but you need to compile the the TDL scrip with the option Calibrate OFF, since the results of the previous calibration are stored in the MS

The aim of this part is to calibrate for differential gains the bright off-center sources and to subtract them properly. The goal is to have the cleanest possible residual. The options and criteria of this new version is completely different between version 1.0 and version 1.1, therefore we will not point out the new features, please refer to the screenshot and to read carefully the info "i". Here is described a different approach from the standard presented in Oleg's and Tony's Tutorials.

SUBTRACT SOURCES: Before to solve for differential gains on the bright off-center sources, we want to subtract from our data the faint sources which are not affected from these pointing errors. We will create a simulated data set that will include the whole sky the model at our disposal excluding the 7 bright sources for which we want to solve for differential gains: UVMODEL=MODEL-7(NEWS###). See screen shot attached 2.1 and 2.2.

SOLVING FOR DIFFERENTIAL GAINS: We a bit f experience should be easy to identify which sources need a differential gain correction before to be subtracted. In our case we used those suggest in Oleg's tutorial. The screenshots 2.3, 2.4 describe the TDL compile options used. The TDL run time Jobs are shown in screenshot 2.5. Before to click Solve for dE you want to open the bookmarks corrected_residual inspector and dE inspector, use screenshots 2.6 and 2.7 to compare the results. Now make a dirty image! (see screen 2.8).

MORE FLYSWATTING: Since few "offensive" sources became visible in our image we decided to solve for differential gains also for these. In order to do that we need to create another UVMODEL=MODEL-11(NEWS###). The procedure in the TDL compile option is the same as before except that the MODEL now include 11 NEWS### sources. The screen 2.9 hows the TDL run time Jobs for which we check off the options "select solvability by subgroup" marking only the newly added sources, since for the others we already solved for it. Find the solutions....Click finally on Make a dirty image and you will be glad to see that your image looks like screen 2.10.