[evlatests] the amplitude glitch in antenna 13

[Ken Sowinski]

I used Rick's test time this afternoon to try to see how
the previously reported amplitude glitch at antenna 13
behaves when phase switching is disabled.  Before doing
anything ea13-l302-2 had to be rebooted because it insisted
on staying in standby mode.

First, an integration time of 1-2/3 seconds (32 WG cycles)
was used.  This is one complete cycle of the Walsh sequence.
During all these observations, which were at X band, no
effect was seen in IFs BD.

Disable phase switching at the antenna, but not the correlator.
There should be no response.  Instead an amplitude response of
about 1/32nd of the expected response was seen with the same
phase as when phase switching was enabled in the integration
at the 10 second tick.

Disable phase switching at the antenna and the correlator.
The responce should be the same as with phase switching enabled
and it was.  There was about a 3 per cent degredation in amplitude
of the integration at the 10 second tick.


An attempt was made with 8 WG cycle integration time with phase
switching disabled at the correlator and a walsh seed of 1 used
at the antenna to obtain the maximum sequency.  We expect no
correlated signal, but instead saw a responce of about 1/10 normal
in the integration at the 10 second tick.

Observations with shorter integration times and phase switching
disabled at the antenna are not easily understood because there
is variable partial correlation during each integration.


It looks like something funny is happening to the DDS phase register
at each 10 second tick; even when phase switching is disabled by
using a walsh seed of zero.  The magnitude of the glitch is consistent
with the phase being wrong for the WG cycle at the ten second tick.
My guess is that the phase register is forced to some value for one
wG cycle at the ten second tick.  It is hard to say whether this
is because of MIB software, DDS FPGA firmware or L305 problems.
Walter's report of the strange and different behavior of BD suggests
that there are two problems, or two L302s responding differently to
an L305 problem.