[daip] VLBA autocorrelations / ACCOR corrections

[Leonia Kogan]

Lincoln wrote:

>I have come across a case in which the results of ACCOR seem to depend on
>source ID for VLBA data, which I understand runs counter to expectation.

>Please see the attached PS file (name: TTT) that shows the results for
>ACCOR (SN1). I direct you to the VLA solutions.  So that you can follow
>the scans, I also attach a TASAV file that has NX/SU files (name:
>TTT.FITS.gz). 

I have checked the cases of raising SN outputs for Y (VLA).
All the cases coincide with the strong sources observed: 0528+134, 0458-20.
The sources have ~3Jy at 2cm. I do not know their flux at 7mm.

I guess that Y is the phased VLA. So its effective area can be large enough 
to see the antenna temperature comparable or maybe even more than the noise 
temperature.

For ideal samplers (digitizers) the mean value of the autocorrelation
spectra does not depend on the noise level at the sampler input.

But the VLBA samplers are not ideal and their output depends on the input 
noise level (on the source if its antenna temperature big. 
That is why you see the effect only at VLA).

The change of the autocorrelation spectra's mean at the level <1% 
which you see at the VLA can be simple explained by this effect.

By the way ACCOR was invented to compensate the samplers error which  
can change from one BBC to another one up to 20%.

I'll answer you second question later

Leonia