[daip] [!11175]: AIPS - KRING dual-pol solutions on single-pol stations (31DEC17 and previous versions)

[Kazunori Akiyama]

KRING dual-pol solutions on single-pol stations (31DEC17 and previous versions)
-------------------------------------------------------------------------------

           Ticket ID: 11175
                 URL: https://help.nrao.edu/staff/index.php?/Tickets/Ticket/View/11175
                Name: Kazunori Akiyama
       Email address: kazu at haystack.mit.edu
             Creator: User
          Department: AIPS Data Reduction
       Staff (Owner): Eric Greisen
                Type: Issue
              Status: Open
            Priority: Default
                 SLA: NRAO E2E
      Template group: Default
             Created: 26 October 2017 06:37 PM
             Updated: 27 October 2017 12:33 AM
           Reply due: 31 October 2017 12:33 AM (4d 0h 0m)
      Resolution due: 22 July 2020 12:00 AM (998d 23h 27m)

Hi Eric,

Thanks for prompt reply.

What we have are data obtained with circular polarization feeds (RR, LL,
RL, LR). Here, I have example plots.
https://www.dropbox.com/sh/ua1gz9whcwewz6y/AABXyNfzwGQ1pKQ_Mzz0ubxwa?dl=0

These ps files show a SN table from KRING and a CL table where this SN
Table was applied with CLCAL.Red points are solutions at R and Blue points
are at L. In this data set, JC (JCMT) and SMAR (a station in SMA) have only
LHCP, so they have only blue points.

You can see that JCMT and SMAR solutions do exist in both SN/CL tables.
However, these stations will disappear in any tasks (POSSM, the next
FRING/KRING, SPLIT/SPLAT) that apply this CL table to data. This issue was
also independently found by other people (cc-ed) on different EHT 2017
data. So, at least this is a common issue for EHT data reduction in AIPS.

We also found that JCMT could survive in some specific observing tracks
when JCMT switch polarizations during observations (but not always!). So,
it empirically seems that single or dual polarizations are a key difference.

I hope that SN tables' weights are the origin of this issue, since this is
the only clear difference in SN tables between KRING and other tasks.

Otherwise, these data may have some issues in UVGET, which seems to run in
tasks that apply a CL table to data.

Do you have any suggestions about what we should check while you are
looking at this issue?

Best, Kazu

On Thu, Oct 26, 2017 at 3:34 PM, Eric Greisen  wrote:

>
>
> Ticket ID: 11175
> Department: AIPS Data Reduction
> Status: Open
>
> ------------------------------
> Dear NRAO User,
> An NRAO Helpdesk staff member has replied to your request for support. You
> can either reply to this email directly or return to the NRAO Helpdesk
> webpage to view the response and post a reply. Please be aware that if you
> reply to this email, there are a few things you should know. Please view
> the knowlegebase article 'Can I respond to my Helpdesk ticket through my
> email?' https://help.nrao.edu/index.php?/News/NewsItem/View/6/
> reply-to-your-helpdesk-tickets-via-email If you choose to return to the
> helpdesk webpage, follow the link below to your ticket.
> Best Regards,
> NRAO Helpdesk
>
> https://help.nrao.edu/index.php?/Tickets/Ticket/View/11175
>
> ------------------------------
> I would think that AIPS would not flag one polarization when the other is
> flagged.
> If you are looking at "formal I" polarization then you must have both or
> if you are
> doing things with cross-hand polarizations then sometimes all 4 are
> required.
> If you want any Stokes value from linear polarizations then all 4 should
> be required
> but the current code does not force that.
>
> I will look at this later today but I do not expect what you report.
>
> Eric Greisen
>



------------------------------------------------------
Staff CP:  https://help.nrao.edu/staff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20171027/b8bedc86/attachment.html>