[daip] New Client Reply - [!CEV-851818]: Help with AIPS UVRFI task

[Nithyanandan Thyagarajan]

New Client Reply: Help with AIPS UVRFI task

Hi Eric,

Thanks for the very helpful reply. I have upgraded my AIPS to the 31DEC13 version. I have also managed to successfully install the MNJ. 

I am having some trouble associated with the UVRFI task. However, it is not as abrupt in nature as the 31DEC11 version. The task seems to go through. 

Before I describe the really troublesome aspect I find with getting the data through the RFI mitigation process, I will walk you briefly through the steps I performed:

The parent data set has 512 channels of which some have been flagged. These flags are such that there are only two sets of contiguously flagged channels that included 16 channels or more (72-97 and 212-228). All the other flagged channels contain only 15 or less contiguous channels. The number 16 is significant because I average 16 channels together later using AVSPC.

UVRFI takes only 'BT' sorted data. Hence, I used UVSRT to sort my data which has already been flagged and calibrated. I apply UVRFI on this sorted data set with opcode='cexp' and certain set of parameters for APARM. Here's the output for UVRFI:

-----------------------------
localh> UVRFI1: Task UVRFI  (release of 31DEC13) begins
localh> UVRFI1: UVGET: doing no flagging this time
localh> UVRFI1: Create F-771145    .UVRFI .   1 (UV)  on disk  1  cno   34
localh> UVRFI1: RNXINI: DATA NOT IN TIME ORDER => NO NX TABLE
localh> UVRFI1: UVAVDO:     211925 Visibility records written
localh> UVRFI1: OUTPUT NX TABLE, IF ANY, IS INCOMPLETE
localh> UVRFI1: Updating tables for IF/FREQID/channel selection
localh> UVRFI1: Copied AN file from vol/cno/vers  1   33   1 to  1   34   1
localh> UVRFI1: Copied FG file from vol/cno/vers  1   33   1 to  1   34   1
localh> UVRFI1: Appears to have ended successfully
localh> UVRFI1: localhos 31DEC13 TST: Cpu=    100.1  Real=    168  IO=      2510
------------------------------

There is a minor complaint about NX table but the task completes without aborting. This creates a UVRFI outclass uv-dataset which I re-sort back to 'TB' order using UVSRT again. This data set does not contain a NX table. I verified that the presence or absence of an NX table makes no difference to the problem I am reporting.

Now, I view this data on a short baseline using SPFLG. It appears fine and similar to the parent data set except perhaps for any minor RFI corrections performed by UVRFI.

The trouble starts the moment I average this RFI mitigated data set using AVSPC. I average 16 channels together using the SUBS option. The output data now contains 32 channels. 

When I examine using SPFLG on the same short baseline, of these 32 averaged channels, many channels have been surprisingly 'blanked'. I say surprising because these blanked channels do not appear at the locations where channels 72-97 and 212-228 would have appeared after averaging and the number of blanked channels are far more in number than expected. There was no flagging intended in any of my procedures above and this extra blanking appears to occur after the AVSPC task.

I also went back and verified that if I bypassed the UVSRT and UVRFI tasks and simply averaged the channels in my parent data set (which has already been flagged and calibrated), I get a vastly different looking output than the one I obtained above. 

I am quite perplexed with this behavior. Could you please point me in the right direction and let me know if I am missing something. 

Thanks,
Nithya

P.S.: I am not sure if it would help to send you a couple of SPFLG screenshots showing these unexpectedly blanked channels. Please let me know if you would like to look at them.

Ticket Details
===================
Ticket ID: CEV-851818
Department: AIPS Data Processing
Priority: Default
Status: Open
Link:  https://help.nrao.edu/staff/index.php?_m=tickets&_a=viewticket&ticketid=2610