AIPS HELP file for VLBARUN in 31DEC22
As of Wed Jan 26 23:05:35 2022
VLBARUN: Applies amplitude and phase calibration to VLBA data !
To run this procedure you MUST type RUN VLBAUTIL and
RUN VLBARUN first to define the procedures in AIPS.
-- Load the data from disk --
DATAIN Disk file name
DOUVCOMP Compress data? Use -1, 0 or 1
OUTNAME File name (name)
OUTDISK Working disk with ample space
-- OR existing file --
INNAME Input file name
INCLASS Input file class
INSEQ Input file sequence number
INDISK Disk number for input file
OPTYPE For CONT, PSEU or SPEC (LINE)
CLINT CL table interval in minutes.
CHREFANT Reference antenna NAME
TIMERANG A good bandpass scan, or use
scan# in TIMERANG(1)+rest=0
INVERS -1.0 46655.0 PC table to use
-1 => don't use Pulse cals
do manual phase cal
CALSOUR List ALL Fring fit & CAL sour
FIRST is bandpass calib
SOURCES ONLY FOR PHASE REFERENCING:
Source *PAIRS* to calibrate
START each pair with phase
referencing calibrator, if
2nd='*': all non-CALSOUR
are phase-referenced to 1st
INFILE Set to DELZN correction file
TECRFILE Local copy of TEC file
'' => download autmatically
'-' => don;t do this corr.
EOPSFILE Local copy of EOP file
'' => download autmatically
'-' => don;t do this corr.
SOLINT Time interval for fringe-fit
IMSIZE Image size for target SOURCES
=-1 for no images
FACTOR CALSOUR IMSIZE (FACTORx128)
DOCLIP not 0 -> use CLIP on all
sources: gives clip levels
> 0 after amp cal
< 0 before amp cal & BPASS
BPASS5 Value for BPASSPRM(5)
DOFLAG > 0 -> use RFLAG on all
sources - clip level is
max (100, DOFLAG) if a
clip limit > 100 Jy is
desired (e.g. masers)
TYSMOLEV > 0 -> clip level in TYSMO
ELEVLIM > 0 -> flag elevation <
INTEXT Input flag commands
SELFCAL > 0 => do selfcal on cals
CALTASK 'FRING', 'CALIB', 'CALIBS'
SMRATES (1) > 0 -> smooth rates in
(2) MWF width hours
(3) clip rates > SMRATES(3)
Hint: when setting directory and e-mail lower-case
letters can be retained by NOT using the close quote.
DOPLOT <= 0 no plots;=1 some plots
>1 huge number of plots
=0.25 if convert known bad
OUTFILE DIRECTORY FOR HTML AND PLOT
FILES: if you want an html
file with plots please set
the dir where the html and
plot files can be put.
OUTTEXT E-MAIL ADDRESS: if you want
an e-mail when the script
is complete, set the
BADDISK Disks to avoid for scratch
Use: VLBARUN is the procedure that uses the VLBA calibration
procedures (VLBAUTIL) to calibrate VLBA data. See the explain
file for a detailed description of the procedure.
NOTE: currently this procedure is intended for simple experiments!
Type RUN VLBAUTIL and RUN VLBARUN to define the VLBARUN
procedures. Do not do this again unless you lose the procedures
somehow. HELP PROCS will show the names of all defined
The procedure is run by typing : VLBARUN
The procedure will check the inputs (as far as it can) to avoid
later interruptions, and then proceed directly, if the data are
already on disk. The procedure will clear the message file when
it starts, so print any messages you need to save. This insures
that the printed message file at the end contains only the
messages from this run.
CLEAN STARTING CONDITIONS
It is best to use one project per user number, as otherwise you
may get into trouble if different projects use the same source
names. t is not recommended to run two pipelines with the same
user number simultaneously, although with different POPS numbers
it should work. Overlapping VLBARUNs at any time during the
processes may also have issues.
Clean up all files resulting from previous attempts of running
the procedure, except for the original UVDATA and perhaps, when
restarting a specific frequency-ID, also not the F(X)POL files.
When restarting from disk, make sure the UV data has no extra
SN, CL, PL files. Note that you may want to keep flag (FG)
tables attached to the input file so that they will be applied
during all stages of the pipeline run. Use the procedure
P_RESTART (which is defined in VLBARUN) to do this. Note that
P_RESTART does not delete FG tables. NEVER DELET FG TABLE 1,
the on-line flags.
If you have any antennas that do not have Tsys or Gain curves in
the data, please load them with ANTAB into TY and GC tables #1.
It is a VERY good idea to run TASAV before you do this to keep
the original TY, GC, BL and also FG tables number 1 in case of
Make sure there is enough disk space available on OUTDISK, about
4 to 5 times the expected compressed UVDATA set. For spectral
line probably you need much more because it will make spectral
line image cubes.
DATAIN......48-character name of the disk file from which to read a
FITS file. It must be in the form
where is the remote computer name, is the
environment variable (logical name) for the disk area in
which the file named is stored. is
usually omitted when the file is local to the current
computer. If DATAIN is not found, the task will try
DATAIN with the character 1 appended.
Beginning 2003-Oct-16 FITLD can read more than one disk
file at a time. In that case, they must all have the
same name except that the last letter(s) are the
sequence number 1 through NCOUNT with no leading zeros.
DATAIN must give the base name only, omitting the
sequence number. Then either
DATAIN = 'MYDATA:file_
DATAIN = 'MYDATA:file_1
can access files in $MYDATA named file_1, file_2,
file_3, etc. Note that this is similar to the pieces of
a FITS pieced UV-table data set, but only one UV-table
data set may be read at a time.
DOUVCOMP....If true (> 0.5) then output data will be compressed
which saves disk space at the expense of losing some
OUTNAME.....The name for the output files. The output files will all
be given the class UVDATA and assigned distinct sequence
numbers as needed.
OUTDISK.....The disk drive number for the output data if loading from
disk. Choose a disk with a large amount of space. If
data are already loaded into AIPS (i.e. you are using
INNAME, INCLASS etc) then OUTDISK is ignored and INDISK
INNAME......Input UV file name (name) that is already loaded.
Note that you CANNOT use both DATAIN/OUTNAME/OUTDISK
INCLASS.....Input UV file name (class) that is already loaded.
INSEQ.......Input UV file name (seq. #) that is already loaded.
INDISK......Disk drive # of input UV file. NOTE: this is also
used as OUTDISK.
OPTYPE......The type of observation and requested output files:
'' = 'CONT' for channels < 65; 'SPEC' otherwise
'CONT' = Plain continuum observations, uses pulse cals
unless INVERS = -1
'PSEU' = Pseudo-continuum observations, uses pulse cals
unless INVERS = -1. The data is NOT averaged
over channels, but the resulting images are
continuum images - used for large FOVs
'SPEC' = Spectral line observations. Manual pulse-cal
is performed. Data is unaveraged and images
are spectral line cubes of the phase referenced
target sources ONLY.
'LINE' = Same as 'SPEC'.
CLINT.......Calibration table interval in minutes. This should
normally be in the range 0.25 to 1.0. If <=0 then
0.25 is assumed.
CHREFANT....Name of reference antenna. If not set, a sensible
antenna in the southwest will be tried (FD, PT, LA or
TIMERANG....Time range on which ALL antennas have good data on the
bandpass calibrator (see below, CALSOUR(1)). Used for
instrumental phase calibration by either the pulse-cals
or manually by running FRING on this scan.
If you know the scan number, you can use the scan number
e.g., set TIMERANGE=66,0 to get the time range
associated with scan 66.
If all 0, then VLBARUN will try to find a good scan by
running BSCAN on the bandpass calibrator.
INVERS......PC table to use, if -1 do manual phase cal
CALSOUR.....All the calibrators should be listed here. All these
sources will be FRING fit. IMPORTANT: the FIRST source
is the source used for instrumental phase and bandpass
calibration. If TIMERANGE was set, it must be a scan on
CALSOUR(1). If you want all sources fringe fitted then
you may use CALSOUR(2)='*' but remember to also specify
CALSOUR(1). These sources will be imaged in (pseudo)
SOURCES.....For phase referencing. List PAIRS of
phase-referencing/target sources. The phase calibrator
should be the first in the pair, i.e., with the odd
index. Each target should be the second in the pair,
i.e., have an even index. All the phase calibrators
must be in the CALSOUR list as well. If your are not
phase-referencing any source, then leave SOURCES blank.
If SOURCES = 'PRCAL', '*' then all non-calibrators are
phase referenced to 'PRCAL'.
There is a maximum of 15 pairs, a total of 30 source
SOLINT......The solution interval (min.) for FRING fit. 0 => 1
minute If SOLINT > Scan/2 (in Multisource) SOLINT =
INFILE......If you want to apply a DELZN corrections file, set
INFILE to the filename. See CLCOR explain file for more
details, the file format is the OPCODE='ATMO' format
(which is the same as DELZN produces).
TECRFILE....Input file. If ''=> file will be automatically
downloaded. If there is a problem with automatic
downloads (this can be a problem with some firewalls)
the file can be can be downloaded manually. See EXPLAIN
VLBATECR for details on how to download the file(s).
The file should contain TEC maps spanning the time
range covered by the input CL table and should be in
IONEX format. If NFILES > 1 then the name MUST be in
standard format CCCCdddC.yyC where C can be any
character, ddd is the day number, and yy is the year.
Also the name given in INFILE must be the first day.
If TECRFILE='-', skip this correction which is also
skipped for frequencies above 12 GHz.
EOPSFILE....Input file. If ''=> file will be automatically
downloaded. If there is a problem with automatic
downloads (this can be a problem with some firewalls)
the file can be can be picked up from:
If EOPSFILE='-', skip this correction because the data
are too new for the USNO file.
IMSIZE......Output target image sizes. Minimum IMSIZE of 512 up to
8192 pixels. Sources listed in CALSOUR are automatically
imaged with a size of 128 by 128 unless FACTOR is set.
Pixel size is automatically set by frequency:
IMSIZE=-1 do not make images.
FACTOR......scaled image size of CALSOURs. Image size of sources
listed in CALSOUR is FACTOR*128. The maximum FACTOR is
DOCLIP......If DOCLIP(1) not 0, clip parallel hand data at
ABS(DOCLIP(1)) Jy and cross-hand data at ABS(DOCLIP(2))
Jy. If DOCLIP(2)=0 no cross-hand clipping is done.
Parallel hand data are flagged when cross-hand data are
and vice versa. The sign of DOCLIP(1) determines when the
CLIP is run. If < 0, then CLIP is run before amplitude
calibration, just before BPASS is run. If > 0, then
CLIP is run after amplitude and bandpass calibration
just before the lengthy FRING step.
BPASS5.....If 0, BPASSPRM(5) = 1. If not zero, BPASSPRM(5)=BPASS5.
HELP BPASS says:
(5) Divide by channel 0 option. *** CHANGED MAY 2010 ***
In general, use 0.1 when using multiple calibration
sources. ICHANSEL default is inner 75 percent.
If ~0 (-0.5 to 0.5,but not 0) the value of "channel
0" is determined and applied (divided into the data)
on a record-by-record basis, either from the average
of channels specified by ICHANSEL or from an external
file (if IN3NAME and IN3CLASS are specified). If
-1 (-1.5 to -0.5), the same is done but the
amplitude of channel 0 is made 1.0 so a phase-only
correction is made. Note that model division is
turned off if either of these options is selected.
If -2 (< -1.5), divide by the average of channels
specified by ICHANSEL (no external file) after the
time averaging. This option allows a normalized
bandpass to be found correcting for any
uncalibrated variations in the continuum. If you
have poor phase stability use 0 or -1, if you have
decent phase stability use -2.
If you have good phase stablility use +1. Note
that division by channel 0 adds biased noise
because the channel-averaged amplitude is by its
very nature biased; the -2 option should be less
biased since the data have been vector averaged in
time first and so are less noisy. The -1 (phase
only) division should avoid the amplitude bias but
then the banddpasses are not normalized. That can
be corrected later with BPASSPRM(10) options.
If > 0.5, do no normalization before determining the
solutions. This requires good stability during
each integration period. BPASSP(10)=3 still allows
a normalized solution which may be even less biased
since it is on an antenna rather than a baseline
Note also that a channel 0 may not be determined
sensibly if the phase changes a lot across the pass
band except by setting ICHANSEL to select a very few
******** Note, if normalization is not done on a
record-by-record basis, users must be very careful
about averaging different calibration sources. Any
channel-dependent flagging will cause the different
sources to be differently weighted in the partly
flagged channels than in the completely unflagged
ones. The shape of the bandpass in this case will
not be correct. Therefore, if channel-dependent
flagging is done and multiple calibrators averaged,
the normalization must be done on a record basis.
DOFLAG......> 0 says to run RFLAG twice for each source, the first
time to set the flagging levels and the second time to
apply them. Plot files are made in each pass but do not
appear in the output html page (if any). If you are
observing a source with strong lines, the value of the
pre-clip in Jy (FPARM(13)) is set to max (100, DOFLAG).
<= 0 -> use any flag tables present with the data file
but do not create any new ones. Warning these RFLAG
flag tables can get large.
TYSMOLEV....> 0 => run TYSMO on the TY table. The input TY values are
clipped between below TYSMOLEV/20 and above TYSMOLEV.
Then the data are median window filtered over a 10 minute
interval and values > TYSMOLEV/10 from the median are
flagged. Then a 15-minute box filter is used to replace
all flagged values with a smoothed value from a nearby
time. Do NOT do this unless you think the TY table needs
regularizing (and many do).
ELEVLIM.....> 0 => Flag all data below ELEVLIM degrees
INTEXT......If not blank, text file with flagging commands to be
read by UVFLG.
Adverb values are entered in a text file with
descriptions of data to be flagged separated by a '/'.
There can be up to 40000 selection criteria in the edit
file. See EXPLAIN UVFLG.
SELFCAL.....> 0 => do self-cal on calibration sources
CALTASK.....' ' or 'FRING' use FRING for the time series and any
'CALIB' use CALIB in both of these
'CALIBS' use FRING for the time series and CALIB
for the self-cal
SMRATES.....(1) > 0 => smooth rates with boxcar of width SMRATES(1)
(2) run a median widow filter of width SMRATES(2) hours
(3) clip rates more than SMRATES(3) in mHz from median
widow filter of widths SMRATES(1)
These are done by task SNSMO. Defaults are 0, SMRATES(1),
DOPLOT......Make diagnostic plots to judge quality of procedure
<=0 => no plots
>0 and < 0.5 => some plots but do not try to make gifs
convert is known not to work
PostScript files will be in OUTFILE
area if set
1 => some plots
2 => lots of plots (could be hundreds)
OUTFILE.....Directory of html and plots files. If you want the
diagnostic plots written to disk and an html file
created to make the plots easier to examine please set
directory here. This is limited to 35 characters
(including final /). Environmental variables are
allowed and must end with a ':'. Mixed environmental
and regular are not allowed, e.g.,
A subdirectory will actually be used so this directory
may contain previous subdirectories. It should not
contain any PostScript files. A text file containing
all of the messages will be saved in the subdirectory
with the name messages.txt.
OUTTEXT.....E-mail address for notifications when script is done or
BADDISK.....A list of disk numbers to be avoided when creating scratch
VLBARUN: Procedure that attempts to calibrate VLBI data blindly.
Documentor: Amy Mioduszewski, email@example.com.
Related Programs: All VLBAUTIL procs, BPASS, SPLIT, IMAGR,..
POSSM, VPLOT and many many more
VLBARUN applies VLBAUTIL procedures to simple VLBI experiments.
VLBARUN attempts to make sensible choices if inputs are left to
default. The only inputs that must be set are:
DATAIN or INNAME (INCLASS, INSEQ and INDISK) -> so the dataset is
CALSOUR(1) -> so VLBARUN knows what calibrator to use as the
bandpass/instrumental phase calibrator
SOURCES -> if the experiment is phase referenced
More detail about defaults
VLBARUN tries to pick sensible defaults if inputs are left blank. Below
describes how it makes these decisions.
OPTYPE='' -> VLBARUN checks the total number of channels and decides from
that whether dataset is continuum (64 or less) or spectral
CLINT=0 -> set to 0.25 minutes
CHREFANT='' -> Chooses the first antenna present from the southwest VLBA
antennas in the order of FD, PT, LA and KP.
TIMERANGE=0 -> Runs the task BSCAN on CALSOUR(1) to select scan with the
highest signal to noise with all the antennas present. If
INVERS>=0 then BSCAN also checks if there are PC table
entries for that scan.
VLBARUN offers the option to create diagnostic plots. If DOPLOT is set
to 0 no plots are made. For 0<DOPLOT<1.5 (i.e. DOPLOT=1) the following plots
-POSSM plot of bandpass solution
-SNPLT of amplitudes in CL table after ACSCL
-SNPLT of amplitudes in CL table after APCAL
-POSSM plots of amp and phase after instrumental phase calibration
-SNPLT of phase, delay, rate and SNR in CL table after FRING
-VPLOT of phase of data for phase reference sources with all calibration
-RFLAG of rms spectrum of time and spectral deviations
-KNTR (contour) plots of the images produced, if images are requested
For DOPLOT>1.5, IN ADDITION to the above plots the following are produced
-SNPLT of amplitudes in SN table after ACSCL
-SNPLT of amplitudes in SN table after ACCOR
-SNPLT of amplitudes in CL table after ACCOR
-SNPLT of amplitudes in SN table after APCAL
-SNPLT of phase, delay, rate and SNR in SN table after FRING
-SNPLT of amplitude in CL table after FRING
Output of plots to disk, creation of html file and e-mail message
VLBARUN will create .gif files on disk in a subdirectory of the
directory defined by OUTFILE of the plots that are requested. It then
makes a vlbarun.html file that includes headings for the plots and at
the bottom a list of the CL tables and what calibration is added to
each. This vlbarun.html file is intended to make examination of the
diagnostic plots easier, so the user can quickly judge whether the
pipeline worked. A link to the message file from the run is also
provided in the vlbarun.html file.
It is STRONGLY recommended that the directory defined by OUTFILE be
empty of PostScript files alsthough it may contain date-stamped
subdirectories. VLBARUN creates and deletes PostScript files in this
directory so it is safer if there aren't any such files for VLBARUN to
potentially clobber or delete. VLBARUN makes a subdirectory in
OUTFILE named by the date and time and the final gif files and
vlbarun.html will reside in the subdirectory. VLBARUN checks if there
is an existing vlbarun.html file and will not run if there is. That
should happen only if you restart VLBARUN within the same minute.
It is also recommended that if OUTFILE is used then an e-mail address
be provided in OUTTEXT. The e-mail will state whether VLBARUN is done
(subject: VLBARUN DONE) or failed (subject: VLBARUN FAILED). The
FAILED message only catches failure do to a problem with how VLBARUN
was set up, if the pipeline fails because a task fails that is not
captured and no e-mail message is sent. If OUTFILE is set, the DONE
message includes the URI to the vlbarun.html file assuming that the
machine you are viewing it on is the "localhost". So if you are on
that machine, you can copy that URI ("file:///etc") into your web
browser and look at the plots. If you are not on the machine where
the plots were created then just copying the file:///... URI will not
CONSIDERATIONS FOR MAC OS X AND OTHER OS THAT MIGHT NOT INCLUDE COMMON
LINUX COMMAND-LINE TOOLS -- the output of plots to display using .gif
and html files requires the command-line command "convert" or "magic".
If you do not have "convert" or "magic" then you cannot use the
OUTFILE input. There are packages you can install on your machine
that will give you access to linux command-line tools like convert.
What does VLBARUN do?
Below are the basic calibration steps that VLBARUN performs. This does not
include all the plotting etc..
0) checks that the inputs are correct as well as it can, sets some
defaults. Prints the INPUTS and runs LISTR with OPTYPE='SCAN' and
places them in the output messages file.
1) run VLBALOAD - if DATAIN is specified, loads the data from a FITS
or FITS-IDI disk file. Merges duplicate table files
Runs VLBAFIX to sort the data if needed and make an
index table. If there are multiple frequency IDs,
VLBARUN stops and tells the user to separate them into
separate data sets.
2) run VLBATECR - corrects ionosphere if the observing frequency is
less than 12 GHz. This procedure uses Global
Positioning System (GPS) models of the electron
content in the ionosphere to correct the dispersive
delays caused by the ionosphere. This correction
is particularly important for phase referencing
experiments at low frequency.
3) run VLBAEOPS - corrects Earth Orientation Parameters (EOP) used in
correlation. VLBI correlators must use
measurements of the EOPs while correlating. These
change slowly with time and therefore the EOPs used
by the correlator must be continually updated.
Even though the EOPs that a correlator uses are the
the most accurate at the time of correlation,
frequently after the correlation the EOPs are
improved. This procedure inserts a correction for
the difference between the EOPs used in correlation
and the current best calculation of the EOPs. This
is particularly important for phase referencing.
4) run UVFLG - If INTEXT is not blank, it is used to add flag
commands to FG table 1.
5) run TYSMO - If TYSMOLEV > 0, apply TYSMO to the TY table making
a corrected second version. TYSMO begins by
flagging all Tsys values less than TYSMOLEV/20 and
greater than TYSMOLEV. It then does a median window
filter with a width of 10 minutes and a clip level
of TYSMOLEV/10. Finally a boxcar of width 15 minutes
computes a smoothed version and all clipped values
are replaced with smoothed values.
6) run UVFLG - If ELEVLIM > 0, run UVFLG to flag all data for which
one of the antennas has a source elevation less than
7) run VLBACCOR - sampler corrections. This procedure corrects the
amplitudes in cross- correlation spectra due to
errors in sampler thresholds using measurements of
8) run VLBAPCOR or VLBAMPCL - correct instrumental delays if there is
more than one IF. These delays or "instrumental
single-band delays" are caused by the passage of
the signal through the electronics of the baseband
converters and can be corrected with either the
pulse cals (VLBAPCOR) or by running FRING on one
9) run CLIP - OPTIONAL: If DOPCLIP(1) < 0, CLIP is run on the data
BEFORE any amplitude calibration using ABS(DOCLIP(1))
as the parallel-hand upper limit.
10) run VLBABPSS - Runs BPASS to calibrate the bandpass response
functions for each antenna.
11) run VLBAAMP - Runs ACSCL which applies the BP table and previous
calibration tables and corrects the
cross-correlation spectra using auto- correlations.
For more information see VLBA Scientific Memo #37
(Craig Walker). Then preforms a-priori amplitude
gain calibration using the gain curves and system
12) run VLBAPANG - correct paralactic angle. The RCP and LCP feeds on
alt-az antennas will rotate in position angle with
respect to the source during the course of the
observation (all VLBA and VLA antennas are alt-az).
Since this rotation is a simple geometric effect,
it can be corrected by adjusting the phases without
looking at the data. You must do this correction
for polarization and phase referencing experiments.
13) run CLCOR - OPTIONAL, if INFILE is not blank will run CLCOR with
OPCODE='ATMO' in order to apply a DELZN output file
specified in INFILE.
14) run CLIP - OPTIONAL: If DOCLIP(1) > 0, then CLIP is run on the
amplitude calibrated data using DOCLIP as the upper
amplitude limits of parallel- and cross-hand data,
15) run FRING - possibly final fringe fit of the data. This step
removes the global frequency- and time-dependent
phase errors. If requested, SNSMO is run to clip
bad rates and then smooth the rates. Only then is
CLCAL run to apply the results to the CL table.
16) run RFLAG - OPTIONAL: runs RFLAG twice on each source, the
first time to get flag levels and the second to apply
them. Plot files are generated.
17) do self-cal - OPTIONAL: runs IMAGR on each CALSOUR source. Then
FRING is run on each using the Cleaned image as a
model. Optionally, SNSMO is run to clip bad rates
and smooth the rates. The resulting solution table
is then applied to the CL table, one source at a
18) run SPLIT - applies final calibration and splits into single
source files. If OPTYPE 'CONT' then frequencies
19) run IMAGR - OPTIONAL, if IMSIZE >= 0 then produce images of all
the sources using the IMAGR autoboxing feature. If
OPTYPE='SPEC', separate image cubes are made for
20) run UVPLT - Make plots of the calibrated UV data (a) averaging
and KNTR all spectral channels and 1/2 minute versus time,
(b) averaging all spectral channels and all times
versus projected baseline length, and (c) no
averaging versus baseline length. Plot a contour
map of the output images.