[evlatests] Phase jitter

[Barry Clark]

I am probably doing something stupid, but....

dataset SOFTWARE.tlo.evla.55119.83645106482.55119.83662796296
consists of observations with three slightly different LO
frequencies:  4488.000, 4488.050, 4488.150, cycling around
changing frequencies every thirty seconds, on 3C273.  Since
there is plenty of SNR, I am looking at phase vs time on a
single channel.  The second-to-second scatter on these plots
seems to be much lower when the LO is set to 4488.050 than
on the other two.  I've looked at a reasonable selection of
channels (near midband) and a few different baselines.

Probably should be checked by somebody who knows what he is
doing.