[evlatests] still some subreflector problems

[Ken Sowinski]

I have seen two examples today where the VLA antennas,
controlled by the CMP were sent rotation commands for
the first scan of a file (rotposcmd is correct), but the
command was not seen at the FRM hardware.  In both cases
the right thing happened at the second scan.  Also in
both cases the files were started long before the
scheduled start time.