[ Basic Info | User Guide ]
Basic Information on atlod
Purpose: Convert an RPFITS file into Miriad uv format.
Categories: data transfer
ATLOD is a MIRIAD task, which converts a uv data-set from the
RPFITS format to Miriad format.
Name of the input RPFITS files. Several names can be given --
wildcard expansion is supported. If a single name is given it
can be a raw tape device name (e.g. /dev/nrst0 in UNIX)
containing several files. In this case, see the NFILES keyword
below. There is no default.
Name of the output Miriad uv data-set. No default.
IF number(s) to select. Default is all IFs. For example,
if you observed with 5 GHz (frequency 1) and 8 GHz (frequency 2)
simultaneously, IF 1 would be the 5 GHz data and IF 2 would
be the 8 GHz data. This now also lets you select zoom bands
using values greater than 2. The freq 1 zoom bands come before
the freq 2 ones.
The rest frequency, in GHz, for line observations. By default,
the value in the RPFITS file is used. Giving a value for the
"restfreq" parameter overrides the RPFITS file value. If you do
set this parameter, you MUST give the same number of values as
the number of IFs written out. A value of 0 is used for a
continuum observation. For example, if you have two IFs, the
first of which is HI, and the second is continuum, use
This gives extra processing options. Several can be given,
separated by comas.
'birdie' For CABB data:
CABB generates self-interference in a number of
channels across the spectrum due to 640 MHz clock
harmonics. These birdies are fixed in channel
number for each CABB configuration. The birdie
option currently knows about the 2048x1MHz continuum
mode and flags the affected channels, 100 band edge
channels on each side, and, at 20 and 13cm, the
unusable parts of the spectrum.
For pre-CABB data:
ATCA self-interference can corrupt channels at
integral multiples of 128 MHz. The birdie option
flags these channels. Additionally, in continuum
(33 channels/128MHz) mode, the birdie option dicards
every second channel, plus some edge channels. The
channels discarded are those most likely affected by
the self-interference. Discarding these channels
does not have a sensitivity penalty, because the
effective channel bandwidth is twice the channel
'reweight' For pre-CABB data: re-weight the lag spectrum to
eliminate the "Gibbs" phenomena in continuum data
(33 ch/128 MHz); ignored for all other data.
'compress' Write output data in compressed format.
'noauto' Discard autocorrelation data. The default is to
copy the autocorrelation data.
'nocross' Discard cross-correlation data. The default is to
copy the cross correlation data.
'relax' Do not flag visibilities based on SYSCAL
information. The default is to flag visibilities if
they have not been preceded by a valid SYSCAL
record, or if the the values in the SYSCAL record
look bad. SYSCAL values are checked for sampler
statistics being within 3% of 17.3%, or 0.5% of
50.0%, that the XY phase is within 10 degrees of its
running median, and that the XY amplitudes are
within 1 Jy or 10% of its running median. The tests
for xy phase and amplitude are skipped for 3mm data
(as there is no noise calibration signal).
'mmrelax' This option is ignored, it is only present for
'unflag' Save any data that is flagged and also keep it's
syscal info. By default ATLOD discards most data
that is flagged.
'opcorr' Correct for atmospheric opacity. This option is
possible for data measured after October 2003.
Because of the way system temperature is measured at
3mm (an "above atmosphere" measurement), it is not
appropriate for 3mm data. This option is silently
ignored for 3mm data. Generally it is only relevant
for 7mm and 12mm observations. It does no harm (and
negligible good) for longer wavelengths.
'samcorr' Correct the pre-Dec93 data for incorrect sampler
statistics. Since December 1993, sampler
corrections are performed online. This option is
silently ignored for data measured after December
'xycorr' Apply the on-line measurements of the XY phase.
This option is silently ignored for 3mm data
measured before October 2007.
'hanning' Hanning smooth spectra and drop every other channel
This option is ignored for 128-MHz, 33-channel data.
'bary' Use the barycentre as the velocity rest frame. The
default is to use the LSR frame.
'noif' Do not map the simultaneous IFs to the IF axis.
By default ATLOD attempts to map the simultaneous
frequencies to the IF axis. This will not be
possible if there are a different number of
polarisations in the different IFs.
'nopflag' If at least one polarisation of a set of 2 or 4
polarimetric spectra are bad, ATLOD normally flags
all of the polarisations. Option nopflag changes
this so that only the nominally bad spectrum is
'hires' Treat bin data as measurements in the high time
resolution mode. The output dataset contains no
bins, but instead appears as data measured with
small cycle times.
'pmps' Undo `poor man's phase switching'. This is an
obscure option that you should not generally use.
'single' Assume input is a single dish RPFITS file (from
Parkes or Mopra). This is usually used together
with option 'relax'.
'caldata' Save visibilities associated with certain system
calibrations. Currently this consists of reference
pointing calibration and "paddle" measurements.
'nocacal' Flag data that atlod suspects is taken during a
CACAL scan. There is potential for error in atlod
determining which data are and are not part of a
cacal scan. Use this with caution.
'nopol' Discard data that is not "parallel hand" Stokes
'rfiflag' Flag channels at frequencies that are known to be
bad. This uses the file rfiflag.txt in the current
directory or the default version in MIRCAT. The
file should contain 2 frequencies per line, the
lower and upper end of the rfi in MHz. Precede
comments with a '#'.
'notsys' Undo the online Tsys correction. Useful if RFI in
the tvchannel range has caused the corrections
to be very noisy. The resulting data will be in raw
counts scaled down by 10^6.
'nopack' Don't pack the two CABB autocorrelation bins into
one bin. The autocorrelation bin 1 and 2 will
contain noise cal OFF and ON data.
The default is to repack the info into bin 1:
XX and YY will contain GTP and SDO data in the real
and imaginary part; XY will contain the OFF data and
YX will contain the ON-OFF data (for XY). This will
allow recalculation of the Tsys and xyphase later.
This gives one or two numbers, being the number of files to
skip, followed by the number of files to process. This is only
useful when the input is a tape device containing multiple
files. The default is 0,1 (i.e. skip none, process 1 file).
NOTE: Using this feature to skip many files on a tape is VERY
inefficient. It is far faster to skip using operating system
commands. When doing this, however, you should be aware is that
every RPFITS files consists of 3 tape files. Thus you will want
to skip three times as many tape files as RPFITS files. For
example, in UNIX, to skip 10 RPFITS files, use
mt -f /dev/nrst0 fsf 30
This gives one or two numbers, being the number of scans to
skip, followed by the number of scans to process. NOTE: This
applies to all files read. The default is to skip none and
process all scans.
This gives the number of frequencies to use per IF for opacity
correction. The default (2) is to use linear interpolation
across the spectrum. Maximum value is 32.
Specify the percentage of edge channels the birdie option will
flag out. The default is 9.8 which will flag about 100 channels
from the bottom and top of a 2049 channel spectrum. This
parameter is only used if the birdie option is specified.
For concatenated spectra the width of a single zoom is used.
Note that noise and artefacts go up quickly towards the band
edge so making this much smaller will not gain you much.
Revision: 1.53, 2016/08/25 23:02:21 UTC
Generated by email@example.com on 26 Aug 2016