[evlatests] evla tests, Aug 1, 2005

[Ken Sowinski]

I will organize this report about topics rather than time.
If there is a large outcry claiming incomprehensibility
I will refrain from doing so in the future.

Delays
While Rob was on his way to the VLA I tried to check delays.  
This was done for L band: IFs A and D were off by a lot and 
were fixed.  The delay for IF D is not much to be trusted 
because the bandpass is so funny.  Delays were not set for 
X band out of laziness and because the phase wind which is 
still there was fairly large making the IF A data not reliable.

Phases
The phase wind, and 10 sec clunk, was still there on IF A,
but not B and D.  It got slightly smaller when switching to
3c273 at 0900 LST.  A while later after the antenna was put 
back in service after being worked on the phasewind had moved
from IF A to B and D.  It certainly smells like software When
that happened I abandoned the thought that something might be
learned by swapping L302's.

Phase switching
Before Rob arrived it was verified that phase switching degrades
the amplitude by almost 1/2 for all IFs, A as well as B and D.
Rob and Jerry reversed the sense of the LVDS driven signal which 
proivudes the heartbeat to the DDSs in the L302.  After that it
was determined that phase switching still degraded the amplitude.
Offsetting the timing by a WG cycle or two in eaither direction 
only made things worse.  We next have to see that the heartbeat
is making it to the innards of the L302 properly.

Downconverter power setpoints
The MIB and executor software to do this was successfully tested
in IF B.  Pete will install the necessary software in all MIBs
which have working output power detectors.  This may help in
keeping the power level to th3 samplers reasonable.

L302
About the time Rob decided to leave the VLA the BD L302 broke.  
the main loop refuses to lock at any of the frequencies we use
leaving only IF A as usable.

DTS modules
Rob and Jerry will bring the IF C DTS module back with them.  Because
of past experiences Rob suspects that the funny bandpass for IF D
is sampler related so we did not try swapping deformatter modules
in an attempt to fix this.  I don't think he brought that one as
well.  It is clear that the autocorrelation bandpass is not quite
right either: it is in fact much flatter than the other IFs.

C band
The remainder of the time was spent looking for C band.  This is when 
it was noticed that the BD L302 had died.  It fringed immediately
at IF A.  Delays were set and the focus peaked at about -0.5 cm.  The
database has not yet been updated.  The response for IF A is much 
weaker than we expect, as it is at all other bands.  There was no phase
wind in IF A.