[daip] Confused about APCAL

[Alan Marscher]

Hi,

Since I and my collaborators reduce a lot of 43 GHz data,
we need to correct for opacity using APCAL.  We investigated
the behavior of APCAL on a dataset in March 1998 using various
settings of the APCAL parameters.  We found that if we set
TREC equal to the nominal values from the gains.key file and
specified an approximate value for the zenith opacity, the
APCAL solutions looked OK but the flux scale ended up being
too low by about 30%.  (This was true for another 5 epochs as
well.) Setting TREC=0 produced a flux scale that was about right. 

However, for epochs starting in Apr 99 (after the automated
calibration was installed), things seem to have changed. Setting
TREC=0 leads to very high fluxes and odd looking time
variability of the opacity corrections. This occurs for data taken
during good weather as well as bad. APCAL seems to be too poorly
constrained. TREC=nominal values seems to work better than it did
before.  Has something changed?

The EXPLAIN file for APCAL isn't very clear about what is done
to the data, so the best parameter settings are not obvious.  The
Cookbook doesn't help much, either.  What are the recommended
settings for APCAL in cases in which the opacity corrections are
fairly high?

One gripe is that APCAL (at least the 15OCT99 version that
we're using) only accepts an ASCII weather file, not the
WX table produced by the automatic calibration.  I don't
know if that's true for the Classic AIPS, since we haven't
installed it owing to the unavailability of a binary-only
installation.

Thanks for looking into this,

Al