[daip] FILLM query

[Eric Greisen]

Elias Brinks writes:
 > When using fillm on a fairly recent dataset (2005/05/07) I get a number
 > of messages saying:
 > 
 > MCCHED...BAD GAIN EXPONENT
 > 
 > there are heaps of them...
 > 
 > I also get an error saying:
 > 
 > CHKORR: ANT 14 CORR 1 DISAGREES WITH FIRST
 > (and the same for corr 2, 3, and 4).
 > 
 > Where can I find more info on these errors, their origin, and  
 > significance?
 > Should I be worried...and if so, why?
 > 
 > I'm using a recent version (few weeks ago installed) of 31DEC07

There really is no place where these are described.  The bad gain
exponent occurs when the antennas are not working and the data are
deleted.  Mostly the data are telling me to take 2.0 to a power like
-1000 or +1000 - way out of line.  I am not 100% confident that the
test might not be a bit too restrictive - but usually it saves
floating point aborts.  The other is a bit that now means something
(data are or are not cross correlations rather than scaled to deciJy)
The task now test that bit that was once used for an antenna that used
an assumed constant nominal sensitivity rather than the measured one.
The number of errors of that sort suggest 1 antenna was not fully
calibrated on line, but all should be more or less okay (unless you
have large changes in system temperature due to spillover, radio
sources, etc.

Eric Greisen