[evlatests] frequency/phase offsets

[Ken Sowinski]

To review, this is the issue of whether the instrument is phase
stable with small chenges in sky frequency.  It was noted that
a phase change given by df*L was seen for a change in frequency
df and propagation time L to an antenna, but that the sign differed
for EVLA and VLA antennas.

The sign is explained because there are different causes.  For
VLA antennas we are seeing the effect of propagation time of the
IF back to the correlator; for the EVLA the fine tuning is done
with the L302 and we are seeing the effect of propagation time
of the LO reference to the antenna.

This model was used to modify the executor for both VLA and EVLa
antennas.  It was found to work for VLA antennas and most EVLA
antennas, but a few EVLA antennas did not behave as expected.
We now think that the aberrant antennas are not properly in sync
at the level of a millisecond or so.  The numerology is consistent:
a frequency offset of 100 Hz and a 1 msec timing error will induce
a phase change of 36 degrees consistent with what we see for the
four antennas.

Further evidence was provided by trying to synchronize antenna 11
the worst of the offenders.  It proved to be very difficult to
synchronize and at one point was off by even more than before.
When in this state the phase change with a 100 Hz offset was even
larger than before.

The four antennas, ea11, ea23, ea24 and ea26 will all be investigated
next week.

This experience suggests a diagnostic for mis-synchronized antennas.
Once the executor adjusts phase for a frequency change we can observe
the same calibrator at two frequencies offset by 100 Hz.  The
expectation is that all phases will remain unchanged.  Any antenans
with a change of phase must not be synchronized correctly.

With this understanding I want to make another test observation with
much larger offsets be sure the algorithm for VLA and EVLA antennas
is reasonable and then update the executor.



The evidence:
This table shows the approximate phase change in degrees for each
of three sky frequency offsets.  Antenna 16 was a stable reference
antenna.

BL    100Hz  200Hz  50Hz
16-24  -23    -60   -18
16-11  -70   -140   -40
16-23  -50   -100   -25
16-26  +20    +30    +7