[daip] AIPS help -message receipt?

[Gary A. Welch]

Eric or whoever,

Just checking that you received the message below, sent 15 March.

Gary Welch


------------------------


Hi Eric,

Following your recent suggestion about the possible origin of the "bull's-eye" 
pattern in some of my data cubes, I looked at a few of the ".LINE" files with 
SPFLG, using DPAR(2)=1.  I could not distinguish between data from days where 
the pattern is present and other days where it is not. 

For the first time, I also used POSSM to look at the same files.  First I tried 
APARM(8)=1 to search for total power data.  None was found in the 3-4 sources I 
tried.  (I assume that "total" = "auto-correlation" as used in your last 
message.)

I then set APARM=0 in POSSM, turned on flagging and calibration, and compared 
plots of IF1 and IF2 data (amplitude vs channel, and phase vs channel).  You 
might recall that only IF2 produces the bull's-eye effect, and only on 2 of 4 
days.  I've looked at 60-90 baselines, in both IFs, and stokes RR (and some LL), 
and at some data from each day.  A clear pattern has (finally) emerged:

Roughly 20% of the IF2 data from "bull's-eye pattern days" shows a systematic 
channel-to-channel phase variation - typically by 40-90 degrees.  The one phase 
calibrator I've looked at shows the same kind of effect, but with a much smaller 
variation.  I emphasize that I'm seeing a channel-to-channel variation in IF2 
and most often stokes RR.  A number of baselines seem to flip between two stable 
phases separated by tens of degrees.  The effect is much less frequently present 
in IF1 data.

I am unable to see anything analogous in data from either IF on days where the 
bull's-eye pattern is absent.

I conclude that the phase variations cause the bull's-eye pattern.  If so, I 
have at last "seen the enemy"!  I am going to use POSSM to check for a possible 
time dependence, but since SPFLG didn't show anything I am not hopeful.  Do you 
have any suggestions?

Regards,

Gary Welch