[daip] negative elevation values: BL149CC

[Yuri Y. Kovalev]

Thank you, Eric, for the quick and useful replay!
As always.

I would risk to share with you my view that this kind of 
information should be distributed to a broader community of VLBA 
users, not just a few people in the cc list of this e-mail.
It would be very important to know for VLBA users that AN table 
format is different for 31DEC08 and 31DEC09 AIPS.
The problem with the pipeline might, probably, be kept hidden 
since it is temporal and, hopefully, will be fixed soon

Happy Thanksgiving,
Yuri

On Fri, 27 Nov 2009, Eric Greisen wrote:

> Yuri Y. Kovalev wrote:
>> Dear colleagues,
>>
>> I am writing you again about AIPS / VLBA data problems during American
>> holidays. There must be some reason why I am so "lucky" during these
>> days. Anyway, here it comes.
>>
>> AIPS ver 31DEC08.
>> MOJAVE 512 Mbps experiment BL149CC.
>> Problem: I got negative elevation values.
>> To track it down I have loaded just one pass (using FITLD) -
>> BL149CC_1_FPOL4IF8MHZ16CH_15.3256GHZ00 (taken from the NRAO archive)
>> than have plotted elevation versus time for the refant 2 (FD). You can
>> see results attached. Any ideas what is wrong and how to fix it?
>>
>> Two comments:
>>
>> I have discovered this problem while doing APCAL with opacity
>> correction. Opacity correction was doing strange things which pushed me
>> to check the elevation values. After discovering the problem I have
>> deleted everything, loaded one pass and the problem was there.
>> No VBGLU involved.
>>
>> To check if this is my local AIPS issue I have loaded into the same AIPS
>> one of my recent BK150 U-band datasets
>> (VLBA_BK150K_FPOL4IF8MHZ16CH_15.3495GHZ03.uvfits) and everything looked
>> fine (positive meaningful elevation values).
>
> You are indeed unlucky.  There are actually two problems - the use of
> the uvfits files means that you are getting 31DEC09 format AN files.
> These do not work well with 31DEC08 and will not get correct elevations
> etc.  There is a second problem as well - I believe that the data you
> are looking at was run through our pipeline when it had an error.  That
> error would cause some of the antennas in the AN file to have the wrong
> By sign - actually those antennas will have the correct sign so far as
> 31DEC08 is concerned and the ones we think are correct will be wrong for
> you.
>
>
> You can try fixing the By by hand with GETTAB and PUTTAB.  The following
> stations should have By positive in 31DEC08:
>
> 'SC','GB','HN','AR','RO','YS','JB','DEF','KNO','DAR','MK2','TAB'
> 'BR','FD','KP','LA','NL','OV','PT','Y','MK','GO'
>
> This probably covers all your stations and then some (I tried looking
> for your data bet there are 1000's of files seemingly...)
>
> A better alternative would be to download the FITS-IDI files for your
> experiment.  The uvfits files are currently corrupted until we fix the
> archive.  So in either 31DEC08 or 31DEC09 your only choice is to
> download FITS-IDI and run FITLD from there plus all of the simple fixup
> procs.
>
> Eric Greisen
>