[evlatests] L302 dropouts explained...

[Walter Brisken]

Matt Luce, Pete Whiteis, and I were tinkering with the pair of L302's in 
the lab and can explain the dropouts now and can make them appear in the 
lab.  Matt and Pete are currently sorting out a fix which could be done 
either in firmware or software.

The problem is that the DDS control register contains both the frequency 
parameters and the phase switch bit.  On the 10 second intervals the 
frequency parameters are changed, but the phase switch bit is not 
preserved (thus the problem).  The phase swich bit will not again be 
changed until the walsh function changes states, which could be any 
integer multiple of 52 ms, but for most antennas just 1 or 2 waveguide 
cycles are affected.

Why does ea24 work without dropouts?  Because the walsh function at the 
critical time is 0 for the particular walsh seed (54) used at this 
antenna.

On the test equipment only the first scan of a file caused dropouts. 
Why?  Because the two L302s with the same walsh function get their 10 
second intervals synced up after the first scan change.

When will the fix be ready?  Perhaps a first fix will be ready before end 
of today.  No hardware changes should be needed.

-Walter, Matt, Pete