; ATLOD ;--------------------------------------------------------------- ;! Reads ATCA data in RPFITS dormat into AIPS ;# TASK UV TAPE ;----------------------------------------------------------------------- ;; Copyright (C) 2009 ;; Associated Universities, Inc. Washington DC, USA. ;; ;; Copyright (C) 1988-2009 ;; Australia Telescope National Facility, CSIRO, Epping, Australia ;; ;; This program is free software; you can redistribute it and/or ;; modify it under the terms of the GNU General Public License as ;; published by the Free Software Foundation; either version 2 of ;; the License, or (at your option) any later version. ;; ;; This program is distributed in the hope that it will be useful, ;; but WITHOUT ANY WARRANTY; without even the implied warranty of ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ;; GNU General Public License for more details. ;; ;; You should have received a copy of the GNU General Public ;; License along with this program; if not, write to the Free ;; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ;; MA 02139, USA. ;; ;; Correspondence concerning AIPS should be addressed as follows: ;; Internet email: aipsmail@nrao.edu. ;; Postal address: AIPS Project Office ;; National Radio Astronomy Observatory ;; 520 Edgemont Road ;; Charlottesville, VA 22903-2475 USA ;----------------------------------------------------------------------- ATLOD LLLLLLLLLLLLUUUUUUUUUUUU CCCCCCCCCCCCCCCCCCCCCCCCCCCCC ATLOD: Task which reads RPFITS format uv data into AIPS. INTAPE 0.0 9.0 Tape drive number (0 -> 1) No REMOTE capability NSKIP * 0.0 Number of tape files to skip (0 -> 0) NFILES * 0.0 1000.0 Number of tape files to load (0 -> 1) BCOUNT * 0.0 First scan to read (0 -> 1) NCOUNT * 0.0 No. scans to read (0 -> all) DATAIN Disk file name OUTNAME Output uv file name (name) blank -> source name OUTCLASS Output uv file name (class) blank -> UVDATA OUTSEQ 0.0 9999.0 Output uv sequence number 0 -> highest unique AND a new file for EACH pass OUTDISK 0.0 9.0 Output disk drive number OPTYPE 'LOAD' load data (default) 'LIST' scan summary 'SUMM' file summary 'SYSC' dump syscal info see CPARM(9) SOURCES * Source selection list TIMERANG * Time range selection FREQSEL * Frequencies to select (MHz) IFSEL * IFs to select CHANSEL 0.0 8192. Channel selection array: (1,n) start channel group n (2,n) end channel group n (3,n) incr. channel group n IFMAP True to map on-line IF chain to AIPS IF axis. NIF 0.0 4.0 Length of IF axis, 0 => ATLOD decides. DOUVCOMP -1.0 1.0 Use compressed data? APARM 1) If > 0, convert to Stokes if < 0, kludge RR,LL,RL,LR if = 0, don't change the polarizations 2) If true, keep flagged data 3) If true, retain autocorrelations 4) Shadowing diameter (m) 5) max NX time gap (min) 0 -> 10 min 6) max NX scan length (min) 0 -> 60 min 7) CL entry interval (min) 0 -> 5 min 8) Number of integrations over which to average Tsys 0 -> 1 See HELP/EXPLAIN 9) Number of integrations over which to average XY phase diff if APARM(1) = 1 0 -> 1 See HELP/EXPLAIN 10) If non-zero drop data on quality of XY phase diff. See HELP/EXPLAIN BPARM Baselines to be rejected, (see HELP) CPARM Further control parameters: 1) System cal. info. print level. See HELP/EXPLAIN 2) If true negate system XY cal phase difference 3) If true, negate phases 4) If true, use XYPHASE array instead of on-line values 5) If true, apply XY phase to the Y gain 6) If 1, report antennas for which sampler stats bad If 2, drop baselines with bad sampler statistics If 3, do 1 and 2 If 4, correct data for bad statistics. 7) Sampler tolerance 8) If true, DATAIN will be read as lower case. This is intended for reading from VAX disks mounted under unix via multinet. 9) If OPTYPE='SYSC' then the 3 text files contain: 1 => XY phase, Tx & Ty 2 => X sampler stats 3 => Y sampler stats 0 => 1 10) If 1, Tsys from IF 1 is applied to IF 2, if APARM(8) is > 0 XYPHASE Enter XY phase difference for each antenna & FQID (degrees) DPARM 1) If > 0, then the scan based source number of all sources is set to 1. 2) If > 0, the spectra are 3-point Hanning smoothed. 3) If > 0 -> set weights to D(3) instead of integration time in units of 15s ---------------------------------------------------------------- ATLOD Type: Task Use: To read and load or list a summary of an RPFITS format uv data file into AIPS. The RPFITS file can be on tape or disk. If OUTNAME, OUTCLASS, and OUTSEQ specify an existing file the data will be appended to it. Adverbs: INTAPE.....Tape drive number. 0 -> 1 Note that ATLOD does not function with the AIPS remote tape system. NSKIP......Number of tape files to skip from current tape position before starting to load data. NFILES.....The number of tape files to load (see OUTSEQ) BCOUNT.....The first scan to read (same in each file). 0 -> 1 NCOUNT.....The number of scans to read (same in each file). 0 -> read all scans from file DATAIN.....Name of an RPFITS format uv data disk file to be loaded into AIPS. If this name is blank, ATLOD assumes that input is to come from tape as specified by INTAPE, and NFILES. If DATAIN is not blank, INTAPE and NCOUNT are ignored. OUTNAME....Output uv file name (name). Standard behavior with default is source name. OUTCLASS...Output uv file name (class). Standard behavior with default is 'UVDATA'. OUTSEQ.....Output uv file name (sequence number). 0 -> highest unique. If NFILES > 1, all the tape files go into one output file. To make different tape files go into different AIPS files, use successive applications of ATLOD and NFILES=1. OUTDISK....Output uv file disk drive number. OPTYPE.....'LOAD' to read the data into an AIPS file (default) 'LIST' to list a summary of each scan 'SUMM' to list a summary of each file (very brief) 'SYSC' to write XY phases and Tsys values into text files in the FITS area SOURCES....List of sources to select. ' ' -> all. A "-" before ANY source name means ALL EXCEPT any source named. TIMERANG...Time range selection, 0 -> all. FREQSEL....Frequencies of data to be selected (MHz), 0 -> all. IFSEL......IFs of data to be selected, 0 -> all. CHANSEL....Channel selection array: 0 => Load all channels Up to 10 groups of channels may be selected. CHANSEL(1,n) selects the begin channel, CHANSEL(2,n) selects the end channel, and CHANSEL(3,n) selects the channel increment for group n. Generally, unselected channels are loaded but flagged. However, if only one group of channels is specified, channels are dropped rather than flagged. IFMAP......If true, then the IF chain number defined by the on-line system is mapped to the AIPS IF axis value one-to-one NIF........Length of IF axis. 0 -> ATLOD chooses DOUVCOMP...If true (DOUVCOMP >= 0) the output data is written in compressed format which can result in a substantial reduction - about a factor of 3 - in disk space needed, but some tasks may not be able to handle this format. Data are stored as 16-bit integers with a different scaling for each visibility spectrum. ATLOD will warn you if there is a loss of accuracy due to some channels (interference?) being much higher than the average. APARM......Control parameters: 1) If > 0, convert XX,YY,XY,YX to Stokes IQUV. If < 0, ignore the actual polarization order and type and call them RR,LL,RL,LR. If = 0, don't change the polarizations. 2) If true, retain data flagged on-line as bad. 3) If true, retain autocorrelations in the output uv file. 4) Antenna diameter to adopt when rejecting data due to shadowing effects. The physical diameter of the AT antennas is 22 meters, so setting APARM(7) smaller than 22 relaxes the shadowing constraint, while making it larger tightens it. The default is 22m. 5) Start a new index entry if a gap longer than APARM(5) minutes occurs in the scan. The default is 10 min., mininum 1 min. 6) Start a new index entry if an uninterrupted section of data on the same source longer than APARM(6) minutes occurs in the scan. The default is 60 min., mininum 1 min. 7) A new CL (calibration) table will be created with empty entries every APARM(7) minutes. The default is 5 min., mininum 1 min. 8) If = -1 Undo the on-line correction and redo with nominal 50 K system If = 0 No additional corrections If > 0 Undo the on-line correction and redo with a Tsys correction averaged over APARM(8) integrations 9) If APARM(1) = 1 then Stokes conversion is requested and the XY phase difference per antenna is used. If A(9) = 0 no averaging of XY phases us done If A(9) > 1 average XY phases for A(9) integrations 10) If A(10) > 0, drop visibilities if one of the antennas has a bad XY phase differences. If A(10) < 0, apply additional absolute XY phase tolerance as well as usual clipping algorithm. See EXPLAIN for clear explanation! BPARM......Baselines to be rejected, specified as 10*I + J where I and J are the antenna numbers, order being irrelevant unless the value is negative. If I is 0 all baselines containing antenna J will be dropped. Examples 12 Drop baselines 1-2 and 2-1 -31 Drop baseline 3-1 (but not 1-3) 4 Drop all baselines involving antenna 4 CPARM......Control parameters for experts only. 1) If = 1 list system calibration information If = 2, you are told when the XY phases or Tsys values appear to have drifted or jumped. If = 3, then the information from 1 + 2 is printed 2) If true negate the system XY phase difference 3) If true, negate the visibilty phase as it is read in (to fix data taken in 1989). 4) If true, use the XYPHASE array rather than the on-line values 5) If true, apply the XY phases to the Y gains. 6) If 1, report antennas for which the sampler statistics are out of range If 2, drop baselines involving antennas with bad sampler statistics If 3, do 1 & 2 If 4, correct data for bad sampler statistics. 7) If > 0, use this as the sampler statistics tolerance instead of the internal default (5% for C(6)=1,2,3 12% for C(6)=4) 8) If true, DATAIN will be read as lower case. This is intended for reading from VAX disks mounted under unix via multinet. 9) If OPTYPE='SYSC' then the text files contain: 1 => XY phase, Tx & Ty 2 => X sampler stats 3 => Y sampler stats 0 => 1 10) Apply Tsys from IF 1 to IF 2 if (A8) > 0 XYPHASE....Array of XY phases (degrees) for each antenna and FREQID. See EXPLAIN DPARM......1) If > 0, all sources are given source number 1. This is for mislabelled single source per scan data taken in March 1992 following intial mosaicing experiments. 2) If > 0, the spectra are 3-point Hanning smoothed 3) By default, the weight of each visibility is set to the integration time in units of 15 sec. If DPARM(3)>0 then the weight for each visibility is set to DPARM(3) ---------------------------------------------------------------- ATLOD: Task to read RPFITS format uv data into AIPS. DOCUMENTERS: Mark Calabretta and Neil Killeen RELATED PROGRAMS: FILLM, FILLR, UVLOD, IMLOD, FITTP The uv file produced by ATLOD is in the multi-source format, that is, it includes the SOURCE and FREQSEL random parameters, and has SU, FQ, and NX extension files associated with it. The SU extension file contains information which associates each source with a set of identifiers (name, ra, dec, etc.), and other information. The FQ extension file records information for each IF used (frequency, bandwidth, sideband, etc). The NX file records the index of visibility records associated with each scan. BCOUNT,NCOUNT Each RPFITS file has multiple scans. These specify the first scan and number of scans to read. If reading from tape, then these numbers are the same for each file. Thus, they probably useful only for single file input. DATAIN If blank, ATLOD assumes that input is to come from tape as specified by INTAPE, and NFILES. Use the AIPS verb AVFIL to skip files on the tape to reach the one of interest. If DATAIN is not blank, INTAPE and NCOUNT are ignored. The DATAIN, when used for disk file input, must be of two forms: 1) 'AREA:FILENAME' where AREA is an upper case environment variable (e.g., %setenv AREA /scratch/data) pointing to the location of the file called FILENAME. The file name must also be upper case. 2) 'FITS:MYDIR/FILENAME' FITS is a pre-defined environment variable and MYDIR is a directory (upper case) below the FITS area. This can be shortened to 'MYDIR/FILENAME' and the FITS area will be assumed as the route. Note if you use DATAIN='SCRATCH/FILENAME' then the files are assumed to reside in /scratch/aips which is symbolically linked to the directory FITS/SCRATCH. 3) We recommend that you simply put your data in the FITS area, and then DATAIN='FILENAME' will work by default. OUTNAME, OUTCLASS, OUTSEQ, OUTDISK Specify the output file name. If they specify a file that already exists, then the new data are appended to it. OPTYPE 'LOAD' means load the specified data into an AIPS UV file 'LIST' resticts ATLOD to reporting summaries of the contents of each scan 'SUMM' restricts ATLOD to reporting summaries of each file 'SYSC' causes ATLOD to write text files in the FITS area The contents depend upon what value you give CPARM(9). By default, you get files containing the XY phases (degrees) for each antenna (XYPHS_UID) and the system temperatures (TSYSX_UID) and TSYSY_UID) where UID is your AIPS user number in decimal. The first column in each file is the UT in days. The second column is the frequency. An integer follows each quantity for each antenna. A zero indicates that value is good, anything else indicates that that value was flagged by the on-line system (e.g. if the telescope was not on source). If you have multiple frequencies in the data, you can use FREQSEL and IFSEL to select only one, otherwise data from all frequencies will be put into the one file. Note however, that the utility program ``pltsys'' provided at the ATNF for plotting the text files produced by the 'SYSC' option, has frequency selection, and it is quicker to load all frequencies with ATLOD and then select the ones you want to plot with ``pltsys''. Use SOURCES to select the desired sources (note that Tsys will vary from source to source). Use TIMERANG to select by time. All other selection parameters are inactive in this mode See CPARM(9) for more details of other possible output files. FREQSEL Allows you to select a number of frequencies. They should be specified in MHz. This apples to simultaneous and time cycled frequencies. They are matched to the actual frequencies in the data with a 1 KHz tolerance. If you select frequencies that come from data with different numbers of channels or polarizations, they cannot all go in the one AIPS file. ATLOD will discard any frequencies that are inconsistent with the attributes of the AIPS uv file. You will be notified as this happens. You will have to run ATLOD as many times as you need output files to deal with this sort of data. IFSEL Allows you to select a number of IFs. The combination of FREQSEL and IFSEL should offer sufficiently flexible selection criteria to deal with all cases. The ATCA produces one or two IFs. Generally, spectral-line observers will use 1 IF, and continuum observers two. It is possible, although not recommended, to observe both IFs at the same frequency. If your correlator setup was such that you had different numbers of channels in the two IFs, then the two IFs would need to go into separate AIPS files. In this case, FREQSEL would be insufficient to discriminate between the two IFs and you would need IFSEL to pick out one or the other. It may also be useful if you are time switching frequencies but only want one IF to be loaded. With just FREQSEL you would need to list all the desired frequencies, with IFSEL you just select the IF number, 1 or 2. Data are selected accroding to FREQSEL *and* IFSEL. CHANSEL Channel selection array: 0 => Load all channels. Up to 10 groups of channels may be selected. CHANSEL(1,n) selects the begin channel, CHANSEL(2,n) selects the end channel, and CHANSEL(3,n) selects the channel increment for group n. If all of CHANSEL(1:3,n) are 0, then no more groups are accepted. Otherwise, the defaults are CHANSEL(1,n)=0 -> 1, CHANSEL(2,n)=0 -> NCHAN, CHANSEL(3,n)=0 -> 1. Generally, unselected channels are LOADED but FLAGGED. An important case in point is when you also set DOUVCOMP = 1 which means the visibilties are stored as scaled 16 bit integers. If you have a strong birdie in a certain channel, you can deselect it with the CHANSEL array. E.g., the bad channel is 10. Set CHANSEL=1,9,1,11,33,1 to load channels 1-9 and 11-33. What this really means is that all channels are loaded but channel 10 is flagged. In addition, channel 10 is not used to work out the scale factors for the visibilty on conversion to 16 bit integers, so that your dynamic range is not compromised in the good channels. Note that if only one group of channels is specified, channels are actually dropped rather than flagged. Thus, CHANSEL=7,29,2 would load only the odd channels from 7 to 29. In the FQ table, you would see that the channel increment has been doubled to reflect this selection. However, the total bandwidth would reflect only the actual channels that were loaded. Really, there should be a channel separation AND a channel increment, but AIPS does not support this. You will also note in the header (IMHEAD) that the frequency reference pixel will also have changed to reflect the new size of the frequency increment. IFMAP AT data may have up to four simultaneous frequencies, but generally you will see only two. Let us call these ``IF chains''. Contemporaneous data is located within the AIPS uv file on the IF axis (if the number of channels and polarizations is equal). When IFMAP is true, (+1) then IF chain 1 will ALWAYS turn up on the IF axis in location 1, and IF chain 2 at location 2. However, you might select, with FREQSEL, only the IF chain 2 frequencies. In this case, you would want them to go to IF axis location 1, otherwise you would waste a lot of disk space by having an empty IF axis location 1, followed by your data in IF axis location 2. When IFMAP is false (-1), then ATLOD will put the next frequecy it encounters in the next available IF axis location, regardless of the on-line IF chain that it came from. I suggest generally you leave IFMAP as false (-1). See the example after the discussion of NIF below for an example where you might set it to true (+1). NIF Normally, ATLOD will work out the length of the IF axis based upon the information present in the first scan header that is consistent with the user's selection crteria. However, it may be you know more than ATLOD and need to tell it the length of the IF axis. Consider the case where you have an observation such as IF chain 1 IF chain 2 Scan 1: 4700 8300 Scan 2: 4600 8000 Let us say you set FREQSEL=4700,4600,8000, thus omitting the 8300 frequency on IF chain 2. Now, you MUST set NIF=2 to get all these frequencies loaded. ATLOD notes that you want one frequency from the first scan, and that it would go to IF axis location 1. But this is the only information it has when it reads the header for the first scan, which is when it must build the attributes for the output file. It doesn't know that the other specified frequencies are is succeeding scans and that they will need an IF axis of length 2. Thus, you must set NIF=2. IF axis location 2 for the first scan will get filled with dummy visibilities of zeros. Note, that if you set FREQSEL=8300,4600,8000 instead, and IFMAP=-1, then the 8300 frequency would be put onto IF axis location 1. However, if IFMAP=1 then it would go to location 2, and location 1 would be zero filled instead. DOUVCOMP For large data sets, you may wish to compress the data format. This saves disk space, but loses dynamic range. The task UVCMP can switch between compressed and uncompressed formats. Uncompressing compressed data does NOT recover that lost dynamic range. APARM(1) If > 0, convert XX,YY,XY,YX polarizations to Stokes IQUV. The format of raw AT data is linear polarizations. At some stage this must be converted to the Stokes parameters IQUV for imaging. Because AIPS does not deal very well with XX,YY,XY,YX data at present, and for other reasons too, we encourage you to use this option and convert to IQUV. Note that the XX and YY gains must be roughly equal (< 10%) for this to work properly if any of your sources are polarized. Check this by loading an unpolarized and strong source (e.g. 1934-63) with no Stokes conversion and compare plots of XX and YY for individual baselines (to do this you need APARM(1)=-1. See below. If < 0, ignore actual polarization order and type and call them RR,LL,RL,LR. This is for cases where you actually want to load the data in their raw form, but in order to get many of the plotting functions in AIPS to work, we need to trick AIPS and call them something it knows how to recognize. Calibrating XX and YY independently assumes that the calibratirs are unpolarized. APARM(1)=0 means load the data as they are with no Stokes conversions or naming tricks. See also APARM(10) for a bit more fiddling about. APARM(2) If > 0 then retain all visibilites that the on-line system flagged as bad. These usually occur for data taken when the system is not yet on source, or for errors detected by the correlator. You would want to be desperate for visibilites to turn this one on. APARM(3) If > 0 then autocorrelations are retained. At present, these contain nothing useful and there is no point to hanging on to them. When they become useful, the default will be change to retain them. APARM(4) The shadowing criterion is currently fairly simple and is a geometric calculation for sources at the pointing centre. For a given shadowed baseline, the shadowed antenna is found, and all visibilities involving that antenna dropped. Note the EW nature of the ATCA is used for the shadowing calculation, so you must disable the shadowing check for LBA data (see below). If = 0 the shadowing is computed for 22 m diameter dishes If > 0 the shadowing is computed for A(4) m diameter dishes If < 0 the shadowing check is disabled Increasing A(4) above 22 provides as more conservative criterion. APARM(5) The index (NX) table contains information about where scans begin and end (times and visibility numbers). This allows rapid access to the data base. A new index record is always started for a source change. If = 0 then begin a new index entry if a gap longer than 10 minutes occurs in the data. If > 0 then then begin a new index entry if a gap longer than A(5) minutes occurs in the data. Minimum allowed is 1 minute. APARM(6) If = 0 then start a new index entry if an uninterrupted section of data on the same source is longer than 60 minutes. If > 0 then start a new index entry if an uninterrupted section of data on the same source is longer than A(6) minutes. Minimum is 1 minute. APARM(7) ATLOD will also create the pristine calibration (CL) table. It consists of 1s (Real part of gain) and 0s (imaginary part of gain). Later, as the calibration procedes, new CL tables will be created by applying corrections to CL table 1. The frequency of CL table entries sets the fundamental time scale on which you can make gain corrections, when you work entirely within multi-source files. That is, when doing self-calibration, it would be pointless having solution intervals smaller than the CL table entry interval. If you split off single source files, the fundamental gain change time scale that you can add is the integration time (10 seconds). Self-cal is often done on time scales of a couple of minutes or maybe less (depends on the S/N), so you should bear this in mind. If = 0 then a new CL table entry will be created every 3 minutes. If > 0 then a new CL table entry will be created every APARM(7) minutes. The minimum is 1 minute. APARM(8) The Tsys correction attempts to correct for gain changes that occur faster than you observe calibrators. It is measured every integration by comparing the switched noise source of known power with the total power for each X and Y feed on each antenna. It is this number that is used to normalize the signals before they are transformed into digital signals in the samplers -- the samplers require input signal levels that are steady. The correlation function is the rescaled by Tsys. Since Tsys is worked out from a single integration it is noisy. It is preferable to apply a less noisy Tsys correction. If APARM(8) = -1, the the on-line correction, if any, is undone, and the Tsys correction is redone with a nominal 50 K system. This is the same as no on-line Tsys correction If APARM(8) = 0, no additional changes are made to the Tsys corrections. If APARM(8) > 1, then APARM(8) integrations will be used to determine the running Tsys mean. Some data are too old for the offline Tsys correction (earlier than 13-Aug-90), you will be warned if this is the case. The algorithm attempts to deal with frequency and source changes (Tsys may jump if the source is bright enough) as well as unexpected drifts or jumps. APARM(9) If you have requested Stokes conversion (APARM(1)=1) then the phase difference between the X and Y feeds on each antenna is used. Essentially, XX and YY must be summed in phase to form Stokes I. These XY phases are noisy and APARM(9) allows an averaged value to be used. If APARM(9) = 0 then no averaging of the XY phases is done and no checks for bad values are made If APARM(9) > 1 then the XY phases are averaged over APARM(9) integrations APARM(10) The XY phase differences for each antenna are measured from a noise diode every integration. Dropouts in normal AT visibilities generally correlate with dropouts in the XY phase difference for at least one of the antennas involved in the baseline. If your source is very weak you cannot detect these dropouts by visual inspection. However, the XY phase difference is always measured. If > 0 then visibilities are dropped when the XY phase difference is bad. The XY phase is deemed bad by comparing it with the running mean and the width of a stack of accumulated phases (see APARM(9)). If < 0, then in addition to first being marked as bad by the algorithm which compares the current point with the stack mean and width (see above), the point is finally only called bad if it is greater than ABS(APARM(10)) DEGREES from the stack mean as well. Note that if you are using the XYPHASE array (CPARM(4)=1), then the values in it are used as the stack mean instead of the on-line values. This provides for a simple means of flagging on the basis of XY phases if you have gone to the trouble of plotting them up. If = 0 then no dropping of visibilities is done. If you have also selected to convert to Stokes, ATLOD will replace what it thinks are bad XY phases with the stack average (see APARM(9) and below). However, visibilities involving the antenna with the bad XY phase are likely to be bad. CPARM(1) If = 1 then lots of information can be printed that is used in the conversion to IQUV. XY phase difference, Tsys and parallactic angles are listed. If = 2, then the user is told when the XY phases or Tsys values appear to have drifted since the last freq or source/freq combination, respectively, was observed. The accumulations of these quantities are reset when this happens. Otherwise, the current accumulations would reflect the wrong values. If = 3, then the information from 1 + 2 is printed CPARM(2) If the conversion to IQUV does not look correct, try turning this on AND come and talk to me (NEBK). CPARM(3) If true, negate the visibilty phase as it is read in. There was a problem with early AT data sets which produced inverted images. The phase had the wrong sign, so you can change it here. Only data from 1989 is likely to be affected. CPARM(4) Use the XYPHASE array rather than the on-line values. CPARM(5) When not converting to Stokes IQUV, you can optionally apply the XY phase to the Y gains. Thus, the YY phases would be rotated to those of XX ******* NB ******** A trap here is that applying XYPHASEs of zero degrees, is different from NOT applying any XYPHASE. Thus, let us say you use the 'SYSC' option to plot up your XYPHASEs, and you see that they are all close to zero. However, because, internally, ATLOD adds pi to the XYPHASEs (to bring them to the same convention as Miriad), applying an XYPHASE of zero is different from not applying them. If you want to apply the XYPHASEs, you should ALWAYS do it, even if they are apparently close to zero. CPARM(6) The sampler statistics for each antenna have well defined ideal values. The correct values are 50% for the zero level statistic. For the positive and negative level statistics, they should be 17.1% for data taken before 21/8/93, and 17.3% thereafter. If the samplers are out of range, it indicates that the data will be scaled incorrectly, although the spectrum may look fine. A tolerance of about 5% is about all that one should tolerate on these statistics. A particular case is that of interference, which may drive the samplers to their rails. This would produce uncalibratable data but would be hard to flag otherwise. If = 1, then when the sampler statistics are out of range, a warning is issued to the user. The default is that the samplers must be within 5% of the correct value. If = 2, then baselines involving antennas with out of range sampler statistics are dropped. The default is that the samplers must be within 5% of the correct value. If = 3, then 1 and 2 above are done. If = 4, then the data are CORRECTED for the error in the sampler statistics when they are within the tolerance of the correct value. The default is that the sampler statistics must be within 12% of the correct value in order that the correction can be made. Any visibilities outside of this range will be dropped. In addition, the correction can only be made of the zero statistic is within 5% of 50% and if the difference of the positive and negative statistic is within 5%. Data outside of this range will again be dropped. Note also that the sampler correction is for small correlation coefficients only. Thus, autocorrelations, if loaded, are not corrected. Similarly, if you have visibility amplitudes of the order of 50 Jy, they should not be corrected either. CPARM(7) The internal setting for the acceptable tolerance in the sampler statistics is 5% if C(6)=1,2 or 3 and 12% if C(6) = 4. If you want a different TOLERANCE, set CPARM(7) to this value. CPARM(8) AIPS normally converts adverbs to upper-case, so reading files with lower-case names is impossible. Files on VAX disks, available to unix via multinet, are seen by default on a unix system as lower-case. If CPARM(8) is set true, ATLOD converts the string in DATAIN back to lower-case, before reading the file. CPARM(9) If you set OPTYPE='SYSC' then information from the SYSCAL group is written into three text files for you to examine it (3 is the maximum number of text files that can be open at one time in AIPS).These files go to the $FITS area. Each file contains 14 columns, the first is the time, the second the frequency, and the next 12 are the quantity and a flag for each antenna. A flagf of 0 means the data are good, a flag > 0 means the on-line system flagged it bad (e.g. not on source for Tsys) If C(9) = 1 the files contain the XY phase differences (XYPHS_UID), the X (TSYSX_UID) and the Y (TSYSY_UID) feed system temperatures. If C(9) = 2 the files contain the X sampler statistics. The negative (XSSNEG_UID), zero (XSSZER_UID) and positive (XSSPOS_UID) statistics are each put into a separate file. If C(9) = 3 the files contain the Y sampler statistics (YSSNEG_UID, YSSZER_UID, YSSPOS_UID). Any other value => 1 UID is your AIPS user number XYPHASE The on-line XY phases for each antenna, are still quite often very poor, especially at 13 and 20 cm. We believe that these phases shoudl be constant over the observation. If you have a better estimate of the XY phase than the on-line value, you can put it in here, and set CPARM(5)=1. You might use OPTYP='SYSC' to write the XY Phases to a text file and examine them. Then input into the XYPHASE array, the mean value for each antenna. If you have more than one frequenciy in the file, then you can input 6 XYPHASEs for the first frequency and then the next 6 for the second frequency. We need to associate the correct XYPHASE with the correct. frequency. The order in which to input the XYPHASEs is the order in which they list in the FQ table. Thus, first 6 phases for IF 1, FREQID 1, then IF 2 FREQID 1, and so on to IF n FREQID 1. Then start on FREQID 2 and so on. Note that if IFMAP=-1 the location of a frequency on the IF axis does not necssarily reflect its on-line IF chain value. You should know what you are doing here in the more complicated cases. See me if unsure (NEBK). You must always specify 6 values if you are going to specify values for more than one frequency. This is true even if antenna 6 is not in use for your observation. SPace for up to 6 frequencies is provided. This would require thirty numbers DPARM(2) If > 0, then the spectra are three point Hanning smoothed. The real and imaginary spectra are smoothed separately. The smoothing occurs before channel selection, so that you can smooth the data, and then drop every second channel say. This is appropriate for spectral-line observations where adjacent channels in the spectrum are intiially independent, but the raw spectrum is still convolved by a SinC function which will cause sidelobes about narrow spectral features. DPARM(3) ATLOD has always written the weights for each visibility as 1.0. However, as of ATLOD version 03/mar/94, and for data taken after 22/feb/04, it now by default sets the weights to the integration time divide by 15 sec. This includes allowing for time losses such as the blank and hold times. If DPARM(3) > 0 then the weights will be set to DPARM(3). So if you wish the weights to still be unity, set DPARM(3)=1 If you wish to concatenate with some old data with weights=1, you should use the REWEIGHT adverb of DBCON accordingly SHOW and TELL are implemented for ATLOD. Type AIPS> SHOW ATLOD to see which adverbs you can change while ATLOD is running. Set their values appropriately, and then type AIPS> TELL ATLOD to inform ATLOD of your intentions. ATLOD receives TELL information in two places. The first is after each specified file has been skipped (NSKIP). Thus you could change NSKIP if you wish. The second is after each scan has been read into AIPS. ATLOD will inform you when it receives a TELL command. Here are the adverbs you can change. OPTELL Operation ('CHAN', 'QUIT', 'ABOR') NSKIP Number of tape files to skip NFILES Number of tape files to load BCOUNT First scan to read NCOUNT No. scans to read SOURCES Source selection list TIMERANG Time range selection FREQSEL Frequencies to select (MHz) IFSEL IFs to select set OPTELL='CHAN' to change them. set OPTELL='QUIT' to have ATLOD finish cleanly at the next available opportunity. VERY useful if you suddenly find you asked ATLOD to read many files from tape, but that you have not selected any data from the last 10 and don't want to wait while it finds this out. set OPTELL='ABOR' to have ATLOD abort The only one that can run you into trouble is IFSEL. Consider loading data where you have set IFSEL=0 and IFMAP=-1. Let us say IF 1 goes to IF axis 1, and IF 2 goes to IF axis 2. Now, with TELL, you change IFSEL=2,0 ATLOD then finds on the next scan that you don't want IF 1 any more, but that you do want IF 2. It sees that IFMAP=-1, which means put the data on the first available IF axis location. Thus, IF 2 will now arrive in IF axis 1. Ths would be a bit of a mess. However, if you set IFMAP=1 in this case, IF 2 would still go to IF axis 2 and all would be well. ----------------------------------------------------------------