[daip] calibration question

[Claire Chandler]

Interesting... when I SPLIT the online CH0 I get

habane> SPLIT1:            Previously flagged   flagged by gain         kept
habane> SPLIT1: Partially               0                   0              0
habane> SPLIT1: Fully                   0                   0         525130

when I SPLIT the LINE data with the same CL table I get

habane> SPLIT1:            Previously flagged   flagged by gain         kept
habane> SPLIT1: Partially               0                   0              0
habane> SPLIT1: Fully                   0               16840         525130

...except that the header of the SPLIT LINE data says there's 508290 
visibilities, so the SPLIT output is lying to me.

So this means it's the BP that's doing the flagging, and this all seems 
consistent.

Claire

On Tue, 27 Mar 2007, Eric Greisen wrote:

> Claire Chandler writes:
>
> > When I run SPLIT and the output tells me that a certain number of
> > visibilities have been "flagged by gain", how do I find out which
> > visibilities those were?  Is there some way of finding this out from the
> > CL table?
>
> The flagging can also come from bandpass calibration if you are doing
> that.  If not, you can try plotting or printing the CL table.  The
> print will show INDE for bad solutions, the plot will omit those which
> means if all IFs/polarizations of an antenna are bad at some time it
> will not be so visible.  You can use color and modes like ALSI or
> overlay graphics channels to try to see what is missing.  The counting
> is done in $APLNOT/DATCAL and is logically somewhat complicated but I
> think it gets things right.
>
> Eric Greisen
>