next up previous contents index
Next: Conversion of single-source files Up: HOW TO COMBINE MULTI-CONFIGURATION Previous: Concatenation

A warning for users of NRAO AIPS

Finally, be aware of the following problem if you are NOT using ATNF AIPS, but are using standard NRAO AIPS. If you combine multi-channel and multi-IF files with DBCON and the frequency increment of the IFs are not all the same, then IFs other than the first may be a problem when you image. With the ATCA, you might have, for example, 13 and 20 cm as the two IFs, and the frequency increments might be 4 MHz and -4 MHz respectively. NRAO AIPS DBCON converts the data to have a frequency axis reference pixel of 1.0 and it scales u and v accordingly, but does not adjust the frequency offsets in the FQ table for each IF to reflect the new reference pixel. If the frequency increments of all the IFs were the same that would not matter. However, for ATCA data where they may differ, this means that if you now attempted to work out the frequency of a given channel from any IF other than the first using the reference pixel, reference frequency, frequency offset and increment stored in the FQ table, you would get the wrong value. This will cause a task such as MX when trying to do multi-frequency synthesis imaging (see § 15) to get the wrong scale for any IF other than the first.

Our solution with ATNF AIPS is to prevent DBCON from making that conversion to reference pixel 1.0. A better solution (and harder to implement) would be to write a correct FQ table on output from DBCON with the offsets appropriately changed. As I have been unable to convince the NRAO AIPS team (what little is left) to do anything about this problem, I am forced to write this stupid section in this manual.


next up previous contents index
Next: Conversion of single-source files Up: HOW TO COMBINE MULTI-CONFIGURATION Previous: Concatenation

nkilleen@atnf.csiro.au