[daip] auto-correlation data in aips

[Robert Braun]

Hi folks,

It would be wonderful if someone could help me out a bit with a small
problem with the way various AIPS tasks treat auto-corr. versus
cross-corr. UV data. After a lot of messing around I've become fairly
convinced that these are not being treated correctly in at least
BPASS, POSSM and SPLIT. The problem is that the BP solutions and their
applications are not self-consistent. For example, if you derive BP
solutions from cross-corr. data, you get an attempt at the _voltage_
bandpass solution, ie. for the IF channel 1R say. When you apply these
solutions to cross-corr. data you form the product of the two different IF
channels, ie. 1R times 2R say. Now, if you attempt to apply these
solutions to auto-cor. data you also would like to get a _power_
BP solution, so you'd like to get 1R times 1R say. This doesn't appear to
be what happens. It looks like you get simply 1R, and not
1R*1R. Similarly, if you derive the BP solutions from the auto-corr. data,
I have the impression that you simply get 1R*1R say, instead of
sqrt(1R*1R), so that it doesn't work well for calibrating cross-corr. data
at all.

Is there anyone left in the AIPS group who is familiar enough with the
relevant tasks that they might look into this, or do I have to start
painfully digging myself into the code?

Cheers,
Robert
________________________________________________________________________

Robert Braun                            |
NFRA                                    | Internet:  "rbraun at nfra.nl"
Postbus 2                               | Phone: (31)-521-595100
7990 AA Dwingeloo                       | Direct: (31)-521-595254
The Netherlands                         | FAX: (31)-521-597332
________________________________________________________________________