[daip] Re: Problems after running MNJ to get the frozen 31DEC04

[Eric Greisen]

Sean Dougherty writes:

 > > How much catching up did the last MNJ require?  Did the COMRPLs and
 > > COMLNKs actually get done properly?
 > >
 > 
 > Looks like the COMRPLs and COMLNKs all ran fine. Certainly the LOG files 
 > you suggested I check all ran fine, with no error messages.
 > 
 > > It has somehow turned your user number into 0 and so cannot find a
 > > message file for that user.  Do you have 31DEC04 as both TST and NEW or
 > > do you have 31DEC05?
 > > 
 > 31DEC04 is both TST and NEW. As yet I haven't downloaded 31DEC05.
 > 
 > 
 > > Try
 > > 
 > > cd $AIPS_ROOT
 > > source LOGIN.CSH     (or . LOGIN.SH for bash)
 > > $CDTST
 > > COMLNK $APGNOT/LISTR
 > 
 > Sadly this did not remedy the problem. I think I'll try a new install - 
 > I know that it likely isn't necessary, but is a useful excersize since I 
 > only do it once a year! I think I'll live on the edge and do with 
 > 31DEC05 in the new install.

    Note that we have a binary installation which might be interesting
to you.  See the web pages.

    I am suspicious that at some time your installation of 31DEC04
thought that it had caught iup on COMRPLs without actually doing them.
This can come from not setting the date right in the initial install
(to match the date of the tar ball not the date of the installation)
even.  The error below from CALIB would appear to be this sort.  If
the uv data single source has 2000 for the epoch then the message
should not say 0.0.  That change was done on June 29.

 > 
 > On a different note, I have a query related to UVFIX. If you run UVFIX 
 > to precess coordinates from B1950 to J2000, the header of the output 
 > multisource file still suggests that the coordinate equinox is 1950.
 > If I split out a source from this precessed multisource file, the header 
 > of the output single-source uv file says Coordinate equinox is 2000.0. 
 > That is good, but is there a reason that the multisource file doesn't 
 > acknowledge that you have precessed the coords to J2000 in its
header?

     You are right - we do not change the header of multi-source
files and should.  We change the source table for them and the header
for single-source files.  I will fix that in 31DEC05.  SPLIT and SPLAT
get the output epoch from the source table.

 > 
 > I appear to have unearthed another, seemingly related problem. If I try 
 > to use a model generated from the newly precessed single-source uv data 
 > file for furhter calibration of the multisource data, CALIB complains 
 > that :
 > BALDRI> CALIB1: SETGDS: IMAGE    1 EPOCH 2000.0 UV EPOCH    0.0
 > BALDRI> CALIB1: USE UVFIX ON THE UV DATA OR EPOSW ON THE IMAGE
 > 
 > OK - run EPOSW on the image (not too many inputs to EPOSW) and:
 > BALDRI> CALIB1: SETGDS: IMAGE    1 EPOCH 1950.0 UV EPOCH    0.0
 > BALDRI> CALIB1: USE UVFIX ON THE UV DATA OR EPOSW ON THE IMAGE
 > 
 > It seems that epoch is not getting set by EPOSW, but also it would seem 
 > that it is not set by UVFIX either. Do I have to set epoch using gethead 
 > and/or puthead?
 > 
 > In addition, I also note that if you set the Besselian epoch of the 
 > input data in UVFIX this is not recorded in the UVFIX history in the HI 
 > table.

   That too I suspect needs fixing...

Thanks,

Eric Greisen


 > 
 > Puzzled.
 > 
 > Sean
 > 
 > 
 > -- 
 > ------------------------------------------------------------------------------
 > Sean M. Dougherty
 > National Research Council
 > Herzberg Institute of Astrophysics		www.drao.nrc.ca/~smd
 > Dominion Radio Astrophysical Observatory        250 490 4359 (direct)
 > PO Box 248, Penticton BC. V2A 6K3	 	250 493 2277 (switchboard)
 > CANADA 
 > 					250 493 7767 (fax)
 > ------------------------------------------------------------------------------