[evlatests] Deep Imaging!

[Vivek Dhawan]

I was working on a comparison of Widar to VLA data, taken in parallel.
As usual, Rick is quick and beat me to the punchline - but I suppose its
worth repeating good news.

The comparison shows that  Widar data look better than the VLA, when
both are treated simalarly (identical treatment is not possible for
continuum VLA data).

Data: 90 min on source 0555+398, 5Jy,  on both arrays.
1. WIDAR: C band, 4 stokes, 4subb x 128MHz. 12 antennas.
2. VLA: C band, 4 stokes, 2IFs of 50MHz continuum. 12 antennas.

Result:

1. Widar:

  a. without any BLCAL tricks:
     Ipol rms=0.31, 0.24, 0.23, 0.22 mJy in the 4 subbands.
     Ipol rms=0.15 mJy averaged over 480MHz. peak/off-source rms ~30,000:1
     Vpol rms=0.17 mJy ""

b. After single BLCAL, time invariant (as per George & Rick.)
    Ipol rms = 27 uJy/bm. Expected is ~18 uJy. Peak/rms > 200K:1
    Upto 5 background sources are visible in a wide image, strongest ~1mJy.


2. VLA with BLCAL (1 time-invariant BLCAL is needed to correct known closure
    errors e.g. VLA-EVLA bandpass mismatch, delay errors of ~1ns, etc).

    Ipol rms=0.19mJy (expect 0.05 mJy for 12 ant)
    Vpol rms=0.17mJy

    peak/off-source rms is 25,000:1. The VLA actually has poorer UV coverage,
    and perhaps there is more can be done to get better images-  but I'm
    quitting while Widar is ahead.



Rick Perley wrote:
>     Some other evidence, favoring George's hypothesis are:
> 
>     a) The required closure corrections for this recent 6cm database are 
> of order 0.3% -- the same level as noted in my previous deep 6cm images, 
> done in different configurations. 
> 
>     b) The closure corrections required for 3.6cm observations are much 
> lower -- typically .03%.   
> 
>     The C-band polarizers are considerably 'leakier' than those at 
> X-band -- the ratio being a factor of a few, as required by the 
> factor-of-ten  difference in the closure corrections. 
> 
>     We should determine if the closure corrections determined on one 
> source can be transferred to another -- if both sources have  negligible 
> polarization (compared to the D-term amplitude), this should be effective. 
>