[daip] aips indxr

[Eric Greisen]

On 05/14/2014 10:53 AM, Cao Hongmin wrote:
> Hi Eric,
>
> 1.
> VLBI multiple source file, after reading the data in aips with fitld,
> the fitld will produce a CL table (CL1) with a default time interval = 1min.
>
> It seems that if we run indxr with aparm (3) =e.g. 1/6, we can change the
> time interval (?)
>
> then what will happen, if we delete the CL1, and run indxr with aparm
> (3) = 1/6,
> does indxr can give us a properly CL table (a new CL1) ?

         If you delete CL table 1, then INDXR will make a new one.  Note 
that FITLD has an adverb that would let you specify this at that stage.
For the VLBA and other telescopes that meet that standard (FITS IDI 
format) of attached tables (MC, IM, etc) INDXR will do as well as FITLD 
with the calibration table.

>
> 2. another is about possm,
>
> after amplitude calibration and manual phase calibration, i got a new CL
> table (e.g. CL9),
> I applied this table and run possm to see the result after calibration,
> one baseline Ef-Sh (see as 1.ps), has a very big amplitude than other
> baseline,
> it is less likely caused by the calibration (amplitude/phase), i then
> checked the
> raw data without applying any cl table, the baseline Ef-Sh appears very
> noisy (see as 2.ps).
>
> what's the possible reason for that, it is because possm just used a
> different scale
> for  different baselines or because something may be wrong during the
> correlation?

Amplitude calibration in VLBI is based on the values in the TY table. 
If some antenna has seriously incorrect values there, then all the 
amplitude calibration for that antenna with be wrong.  The DOCAL=-1 plot 
for that baseline is extremely noisy, applying only the CL table cannot 
make it as much better as would appear in the DOCAL=1 plot.  I suspect 
that you have very bad data in some of the times for that baseline and 
you should try examining it in some other way, such as with SPFLG. 
POSSM uses only the information in the CL  table, just as SPFLG, UVPLT, 
SPLIT, etc will use.  I suspect that POSSM is confusing me by using 
milli and micro Jy in the uncalibrated plot and Jy and milli Jy in the 
calibrated one.  If the phases of EF-SH are all over the place, then any 
time averaging will cause the vector average signal to go near zero and 
appear very noisy.

Eric Greisen