All LBA data correlated before 1 January 2010 have amplitude scaling applied at the correlator based on the nominal SEFD for each telescope. This means that the amplitudes on each baseline from the correlator are roughly in Jy. In order to apply measured Tsys corrections, this factor needs to be “undone” in post-correlation processing. This can be easily done in AIPS task
ANTAB by inserting the nominal SEFD applied to each antenna into the denominator of the
FT keyword value in the
TSYS group in the calibration text file (see the explanation of
ANTAB in AIPS).
For all data correlated at Curtin in 2008-2009, the nominal SEFD values used were as listed in the table below (from the old
vex2config.pl). Note that for the ATCA tied array, it was assumed that 5 antennas were tied, which may not have been the case (see the observers wiki or logs). If less than 5 antennas were used in the tied array, then the nominal scaling factor applied is incorrect. The value applied for each participating telescope is listed in the correlator input file (TSYS keyword) for each experiment, and ideally this should be checked rather than assuming that the values in the table below were used.
Another way to adjust the nominal amplitude scaling (other than in
ANTAB as above) is to use AIPS task
OPCODE = 'GAIN'). Run once for each antenna as required, specifying
CLCORPRM(1) to give the appropriate scaling factor: note that the gain correction factor in
CLCOR corresponds to units of square root of antenna SEFD (Jy).
Hypothetical example: A 22GHz LBA dataset correlated in 2009 has 'TSYS: 106' for ATCA listed in the correlator input file. This assumes 5 antennas were tied, but according to the observer's log on the wiki, only two ATCA antennas were included in the tied array. Therefore the a priori SEFD should have been 265 Jy (=530/2).
If the measured Tsys log is available (pre-CABB), this will be correct for the number of antennas in the tied array. The measured Tsys can be converted to the
ANTAB input format and used to calibrate the amplitudes. The value for
FT should be divided by 106 to undo the nominal scaling applied at the correlator.
If the measured Tsys is not available (which is the case since the CABB correlator has been used at the ATCA), then either
ANTAB input can be generated with nominal Tsys values, or
alternatively, the gain could be adjusted to the (correct) nominal value using task
CLCORPARM = 1.58, 0 (= sqrt(5/2)).
|Telescope||Nominal SEFD (Jy) in each frequency range (MHz)|