[daip] CLCAL bug

[Eric Greisen]

Zsolt Paragi writes:

 > We found a bug in CLCAL, which seriously affects
 > experiments in which much Tsys data is missing.
 > Our problem is that there are timeranges where
 > Tsys measurements are lost at some EVN telescopes.
 > The script that processes the station logs places
 > "-2" into the input file for ANTAB (this is 999.9
 > in the case of the VLBA). APCAL creates an SN table
 > in which these instances are indicated with "INDE".
 > Unfortunately, CLCAL does not interpolate over
 > the timeranges with "INDE" entry.
 > 
 > This is probably not serious for the VLBA, because
 > missint Tsys data are not so frequent. But for
 > us it is a problem since long timeranges may be
 > flagged in the data for some telescopes. We tried
 > to play with the ANTAB input file replacing "-2"
 > with other numbers but it did not help. A short
 > term solution would be to modify the ANTAB files
 > for EVN users: we could replace "-2"s with fake
 > Tsys data, e.g. the last valid measurement. But
 > it would be best if CLCAL could do the job with
 > interpolating over these timeranges.
 > 

I am doing an excavation on my Inbox and discovered this old,
well-founded complaint.  I hope that you have noticed that we did do
an overhaul of CLCAL - hard to miss since the adverbs changed a lot.
The main work was done in May 2003.  The smoothing options now allow
for SN data to be time smoothed but only replacing the blanked values
with smoothed values which would solve the problem with INDEfs causing
flagging.  Further, I just fixed the task so that it actually will
merge SN records that are for the same time/antenna/FQID etc replacing
blanked values with good ones as much as possible.  This arose when
APCAL was run N times for N IFs so as to allow for different receiver
temperatures.

Sorry to be so slow in replying.

Eric Greisen