[evlatests] VLBI @ VLA update

[Vivek Dhawan]

Recent tests (TY012, 013, 014) are summarized. All were useful, but fell
short of the desired 2Gbps with fully phased array.

So far:

Single dish Y1 out of WIDAR, writing VDIF packets to MK5C, works and can
be correlated.

     . 4 streams (2 subband pairs, AC + BD) have been written.

     . 4x16  MHz, using the legacy narrowband BBC and MK5B on the VLBA.

     . 4x128 MHz, using the digital downconverter (DDC) + MK5C on the VLBA.

The TY014 test used a version of the DDC that rendered LCP on all VLBA
stations weaker than RCP. Thus, 2Gbps have been recorded and correlated,
for Y1, but R and L fringes had different amplitudes. An updated DDC has
been deployed since then, and a VLBA-only test (TD034) has already been
recorded, but not yet correlated.

In addition:

Y27 has been phased successfully, as seen in VLA data (which is not news),
and the resulting 2Gbps VDIF packets written to MK5C. However, when phasing
is turned on, the lag-frame packets find their way into the VDIF stream,
rendering the data uncorrelatable by DIFX. This is mostly understood now.

TY015 is planned for next Tuesday, with known hurdles removed.