[evlatests] doppler tracking

[Gustaaf van Moorsel]

Earlier tests had shown that whereas fixed frequency observing
leads to usable data on VLA - EVLA baselines, Doppler tracking
sometimes (but not always) gives bad data.  A likely culprit
was pointed out by Barry: observe files created by JObserve
always have Fluke frequencies ending in a 0, whereas sky freq-
uencies (and therefore Flukes) calculated by the new system can
have any number in the last (Hertz) field.

I did a quick test run this morning to check to what extent
failure/success on crossed VLA/EVLA baselines depends on the
value in the Hertz field.

Preliminary result:

1 - Any even number of Hertz (including, of course, '0') in the
     flukes leads to excellent data on all baselines

2 - Any odd number of Hertz in the flukes leads to no fringes on
     crossed baselines

Gustaaf