[evlatests] VLA timing

[Barry Clark]

According to my reading of the code, within a ten second cycle, early
in the cycle, at about 0.5 sec TAI (7.5 sec UTC), the phase commands 
for the cycle are calculated and sent to the CMP.  There are eight of
these commands send, with execution times of 10.104, 11.354, 12.604, 
etc., TAI.  The phases are calculated for times of 11.125, 12.250, 13.750,
etc. TAI.

In my interpretation of events, these sit in the CMP command queue for 
slightly under 10 seconds before the first is executed.  At approximately 
the execution time in the command, the values in the command are copied 
into the phase command buffer in the CMP.  At waveguide cycles beginning
at 10.990, 11.042, 11.093, 11.146 the phase commands are sent to the L7.
At the waveguide cycle beginning 11.198 the phase command strobe is sent
to the L7, which should result in these commands being used beginning at
11.125 seconds, the nominal time.

Perhaps Hichem can review what is actually happening in the CMP and tell
us if he agrees with this sequence.