Date - 2001.12.07 Tester - S.T. Myers (AOC) Platform - Linux (sandrock) Note: bugs are denoted by lines with prefix: >>>BUG: and queries are denoted by lines with prefix: >>>QUERY: while comments are prefixed by: >>>COMMENT: Some enhancement requests are given after >>>REQUEST: ------------------------------------------------------------------------------- GOAL: Sort out polarization calibration problems in K and Q band data. 1. Date - 2001.12.07 Version - Daily (1.7 build 045) Directory - /home/sandrock/smyers/Testing/2001-12-07/ Data - 20010204AK.UVF Script - 2001-12-07-myers.g par.pwdir := '/home/sandrock/smyers/Testing/2001-12-07/'; par.prefix := '20010204.kband.'; # Prefix for files generated par.fitsf := '20010204AK.UVF'; # Name of UVFITS file par.msf := '20010204AK.ms'; # Name of ms par.abscal := '1331+305'; # Name of primary flux cal par.absid := 6; # FIELD_ID of primary flux cal par.angcal := '1331+305'; # Polarization angle calibrator par.absang := 66.00; # R-L Phase Difference of PCal par.polcal := ['1159+292']; # Name of d-term calibrator par.polid :=1; # FIELD_ID of d-term calibrator par.refsrc := ['0713+438', '0854+201', '0927+390', '1159+292', '1229+020', '1256-057', '1310+323']; # Names of secondary sources par.nfld := 8; # Number of targets par.fieldlist := ['0713+438', '0854+201', '0927+390', '1159+292', '1229+020', '1256-057', '1310+323', '1331+305']; # Names of targets par.gsolint := 60.0; # solution interval (secs) for gain par.grefant := 8; # reference antenna for gain sols par.clniter := 200; # Number of clean iterations desired par.clcell := '0.02arcsec'; # image cell size (Kband) - makemsf(par) MeasurementSet Name: 20010204AK.ms MS Version 2 Observer: AC524 Project: Observation: VLA(29 antennas) Data records: 160234 Total integration time = 8020 seconds Observed from 04-Feb-2001/08:27:30 to 04-Feb-2001/10:41:10 Fields: 8 ID Name Right Ascension Declination Epoch 1 1159+292 11:59:31.83 +29.14.43.83 J2000 2 0713+438 07:13:38.16 +43.49.17.20 J2000 3 0927+390 09:27:03.01 +39.02.20.85 J2000 4 0854+201 08:54:48.87 +20.06.30.64 J2000 5 1310+323 13:10:28.66 +32.20.43.78 J2000 6 1331+305 13:31:08.29 +30.30.32.96 J2000 7 1229+020 12:29:06.70 +02.03.08.60 J2000 8 1256-057 12:56:11.17 -05.47.21.52 J2000 Data descriptions: 2 (2 spectral windows and 1 polarization setups) ID Ref.Freq #Chans Resolution TotalBW Correlations 1 22485.1MHz 1 50000 kHz 50000 kHz RR RL LR LL 2 22435.1MHz 1 50000 kHz 50000 kHz RR RL LR LL Antennas: 26 ID= 1-7: 1=VLA:N72, 2=VLA:N16, 3=VLA:N24, 4=VLA:N56, 6=VLA:E8, 7=VLA:E16, ID= 8-12: 8=VLA:N8, 9=VLA:E28, 10=VLA:W24, 11=VLA:W32, 12=VLA:E20, ID= 13-18: 13=VLA:W36, 14=VLA:E4, 16=VLA:W12, 17=VLA:E36, 18=VLA:N40, ID= 19-23: 19=VLA:N48, 20=VLA:E32, 21=VLA:W4, 22=VLA:W8, 23=VLA:W16, ID= 24-26: 24=VLA:W20, 25=VLA:N32, 26=VLA:W28 Tables(rows): (-1 = table absent) MAIN(160234) ANTENNA(29) DATA_DESCRIPTION(2) DOPPLER(-1) FEED(29) FIELD(8) FLAG_CMD(0) FREQ_OFFSET(-1) HISTORY(2534) OBSERVATION(1) POINTING(348) POLARIZATION(1) PROCESSOR(0) SOURCE(-1) SPECTRAL_WINDOW(2) STATE(0) SYSCAL(-1) WEATHER(-1) Try msplot (from GUI) on ms file, "display as image" - warning, event glishtk.fail () dropped LocalExec::SetStatus: abnormal child termination for /home/cluster/aips++/daily/linux_gnu/bin/glishtk had to kill glish. Restart aips++ and try again - warning, event glishtk.fail () dropped LocalExec::SetStatus: abnormal child termination for /home/cluster/aips++/daily/linux_gnu/bin/glishtk again. 2. Version - Weekly (1.7 build 033) This seemed to work OK. I will assume there is some incomplete fix in daily. >>>BUG: glistk crash in daily msplot "display as image" in daily Examine data amplitudes. Had a hard time getting a reasonable transfer function to see the data. Try fiddling under "Adjust" panel. Note, tried the hue,saturation,value options under the "Color Mode" and they just left a blank picture no matter what I did - what are they supposed to do? Try to find out from User Ref Guide - it is nearly impossible to find out the GUI functions on the msplot page. Eventually went to the viewer-tool page, then up to the Viewer-Module page, then to the gui description. This is not a good organization - should be linked from the msplot page as a likely query a user might want to make. Eventually found that by using Greyscale1 and a -2.5 scale power, could get something useable. But this seems much more fiddly than in AIPS - why? Maybe use a log-scale? Try to kill all of AN13 and AN18 from GUI (could use killant function from my script). It looks like the spectral windows are alternating integrations - this doesnt seem like a very good thing to do! Why not add these as more planes (e.g. 5-8)? I want to be able to blast all the IFs and Pols for a given antenna between two timeranges (or all times) as in TVFLG. I dont see how to do this. I snagged Ralph and he pointed that the ALL "Antennas" button did the first antenna of the selection versus all others (duh, the popup said this but I missed it). Select "Antennas","Correlations","Spectral Windows", "Timerange". Use bombsight and double-click on a point for AN13:ANxx. Flag. This worked! However, it left me zoomed in to some really strange thing (like 2 ants). Had to unzoom twice. Note this happened when I went back in to msplot after leaving earlier. I am not that thrilled on having to double-click in the viewer to get selection of something. I would prefer that clicking on the relevant panel to the left of the window (like the bombsight) assigned the 3 mouse buttons to THAT task - I find little use with assigning the 3 buttons to some different task. Then the left button could move the point, the middle could select it, and the right could cancel. >>>REQUEST: The mouse buttons in the viewer should reflect a single task selected from the left menu. Note that although msplot is greatly improved, it is still not useable for my purposes as it lacks from the display 1. The ability to display the difference between an integration (or pixel) and the average of adjacent times (not baselines). 2. The IFs are interleaved instead of different planes or grouped vertically 3. If the dataset will not fit in the screen, then it will average in the two axes (even across interferometer baselines) rather than providing scrolling. Point (1) is the most serious - there are often single integrations that are 10% dropouts in amplitude that are hard to find. Also, being able to do difference in phase is very useful (note that plotting absolute phase is useless). Done with msplot. 3. Try mapping - automap(par) 1331+305 spwid= 1 [I=2.517, Q=0, U=0, V=0] Jy, (Perley-Taylor 95) 1331+305 spwid= 2 [I=2.521, Q=0, U=0, V=0] Jy, (Perley-Taylor 95) Flux density for 0713+438 (spw=1) is: 0.534 +/- 0.001 Jy Flux density for 0713+438 (spw=2) is: 0.536 +/- 0.001 Jy Flux density for 0854+201 (spw=1) is: 2.769 +/- 0.003 Jy Flux density for 0854+201 (spw=2) is: 2.773 +/- 0.003 Jy Flux density for 0927+390 (spw=1) is: 8.226 +/- 0.012 Jy Flux density for 0927+390 (spw=2) is: 8.251 +/- 0.012 Jy Flux density for 1159+292 (spw=1) is: 1.364 +/- 0.001 Jy Flux density for 1159+292 (spw=2) is: 1.366 +/- 0.001 Jy Flux density for 1229+020 (spw=1) is: 14.771 +/- 0.020 Jy Flux density for 1229+020 (spw=2) is: 14.789 +/- 0.019 Jy Flux density for 1256-057 (spw=1) is: 22.816 +/- 0.030 Jy Flux density for 1256-057 (spw=2) is: 22.825 +/- 0.031 Jy Flux density for 1310+323 (spw=1) is: 1.219 +/- 0.001 Jy Flux density for 1310+323 (spw=2) is: 1.221 +/- 0.001 Jy 0713+438 IF1 I= 0.534 P= 0.005 chi= -41.54 0713+438 IF2 I= 0.535 P= 0.004 chi= -89.03 0854+201 IF1 I= 2.767 P= 0.127 chi= -90.64 0854+201 IF2 I= 2.771 P= 0.124 chi= -93.50 0927+390 IF1 I= 8.216 P= 0.115 chi= 131.19 0927+390 IF2 I= 8.243 P= 0.122 chi= 135.91 1159+292 IF1 I= 1.363 P= 0.020 chi= -82.81 1159+292 IF2 I= 1.365 P= 0.020 chi= -82.66 1229+020 IF1 I= 14.744 P= 0.976 chi= -89.95 1229+020 IF2 I= 14.770 P= 0.981 chi= -87.36 1256-057 IF1 I= 22.774 P= 2.502 chi= 102.12 1256-057 IF2 I= 22.788 P= 2.484 chi= 102.06 1310+323 IF1 I= 1.219 P= 0.031 chi= -44.61 1310+323 IF2 I= 1.221 P= 0.031 chi= -49.95 1331+305 IF1 I= 2.518 P= 0.322 chi= 66.00 1331+305 IF2 I= 2.522 P= 0.323 chi= 66.00 Versus AIPS results: I Peak mJy P Peak mJy Pol % RLPD deg -------------- -------------- ----------- -------------- 0713+438 511.29 ± 0.90 1.41 ± 0.23 0.28 ± 0.04 31.76 ± 7.78 0854+201 2668.92 ± 0.65 115.23 ± 1.90 4.32 ± 0.07 -88.22 ± 0.92 0927+390 8181.94 ± 2.52 142.57 ± 7.00 1.74 ± 0.08 124.92 ± 1.45 1159+292 1377.02 ± 1.44 15.70 ± 0.64 1.14 ± 0.05 -100.76 ± 1.93 1229+020 14218.67 ± 12.01 858.19 ± 12.05 6.04 ± 0.08 -86.88 ± 1.81 1256-057 21928.05 ± 34.34 2544.29 ± 1.50 11.60 ± 0.01 97.89 ± 0.21 1310+323 1186.02 ± 0.02 23.72 ± 0.42 2.00 ± 0.04 -46.76 ± 0.94 1331+305 2408.60 ± 0.45 324.17 ± 1.42 13.46 ± 0.06 66.00 ± 0.00 The agreement between IFs is very good, somewhat better than in the 2001-03-20-myers.txt results, but there do seem to be significant offsets versus the AIPS results though only a few degrees. Note however that 1159+292, which was the polarization calibrator, is signficantly off (~20 degrees). 4. I note in the VLA Chapter of Getting Results that they start the d-term calibration from [I,0,0,0] rather than [I,Q,U,0] like I did. Lets try this (change script). Blow away old tables images etc. but start from the MS - I hope this will work that way (will the data columns reset ok?). No - this clearly did not work. I need to reset the appropriate columnes (CORRECTED_DATA and MODEL_DATA?). In the process of looking at John Benson's pages I just noticed that the links at the top to MS Functions, Calibrater Functions and Imager Functions which are supposed to go to appropriate parts of the User Reference need to be updated as the node #s seem to have changed. I dont like the way they name pages as there are no stable link names. >>>BUG: There should be stable link names to documentation rather than stuff like http://aips2.nrao.edu/daily/docs/user/General/node354.html so that (especially outside) links can be more stable. After talking to George, I think the problem was in setjy which I did only on 1331+305 after learning the columns default to 1 at start. I will do setjy on all source as I did before I made that change. This worked - results: 0713+438 IF1 I= 0.534 P= 0.005 chi= -41.53 0713+438 IF2 I= 0.535 P= 0.004 chi= -88.26 0854+201 IF1 I= 2.768 P= 0.133 chi= -86.97 0854+201 IF2 I= 2.771 P= 0.130 chi= -89.59 0927+390 IF1 I= 8.215 P= 0.127 chi= 118.72 0927+390 IF2 I= 8.243 P= 0.127 chi= 123.54 1159+292 IF1 I= 1.363 P= 0.018 chi= -83.45 1159+292 IF2 I= 1.365 P= 0.018 chi= -82.33 1229+020 IF1 I= 14.743 P= 0.979 chi= -92.93 1229+020 IF2 I= 14.766 P= 0.974 chi= -90.32 1256-057 IF1 I= 22.774 P= 2.509 chi= 103.66 1256-057 IF2 I= 22.790 P= 2.505 chi= 104.25 1310+323 IF1 I= 1.219 P= 0.030 chi= -36.63 1310+323 IF2 I= 1.221 P= 0.031 chi= -35.19 1331+305 IF1 I= 2.519 P= 0.331 chi= 66.00 1331+305 IF2 I= 2.524 P= 0.332 chi= 66.00 This certainly gave a different set of answers. Note 1310+323 has the IFs closer, but they are farther off in angle than what AIPS gave. But 0927+390 and 0854+201 are closer. And 1159+292 is still off by 20 degrees. I do not understand what is going on here. 5. Will try to develop a glish script to do the atmospheric opacity and antenna gain curve correction. Looks like it will be straightforward (though probably inefficient) to do this using measures. George provided me the following recipe destined for the recipes archive: # Example 7: measures: #--------------------- include 'measures.g' print vlapos:=dm.observatory('VLA'); print dm.measure(vlapos,'WGS84') # convert to geodetic dm.doframe(vlapos) print tim:=dm.epoch('utc','today') ; dm.doframe(tim) print srcdir:=dm.source('0234+285') print 'AZ =', dq.convert( dm.measure(srcdir,'AZEL').m0,'deg') print 'EL =', dq.convert( dm.measure(srcdir,'AZEL').m1,'deg') # should use getvalue (not .m0, etc.) above! Next step is to turn this into a function for this. ------------------------------------------------------------------------------ Testing logs archived at http://www.aoc.nrao.edu/~smyers/aips++/testlogs/ This file can be found as 2001-12-07-myers.txt. The scripts can be found as 2001-12-07-myers.g