[evlatests] tests, Oct 25 2005

[Ken Sowinski]

                            EVLA Tests
                         October 25, 2005

This was a day of some confusion and a little progress.

The first few minutes were given to Chris Carilli for a 200 MHz
system test.

Antenna 14 was unavailable because of cryo work which started
just before SW time.  The antenna was back in service shortly
and the L and X band systems were warm, but cooling.  When 14
was returned it did not work at all, just as Rick has reported.
The problem was traced to the FRM contrller which had been left
in local control.  The hardware crew fixed the 14B sampler by 
cleaning lots of optical connectors and cycling power.  The
14D sampler was broken in the process and was cured by cycling
power.  This left 14 in a usable state.

By this time 16 was out for C band card cage work.

Since 16 was gone I tried, unsuccessfully, to get Q band working 
on 14.  The power levels were much to low and setting all 
attenuators to zero provided just enough signal to the samplers
and gave a hint of a SD voltage at the T5, and there was no hint
of any fringe power.  A consultation with Bob showed that the Q band 
receiver looked to be hot.  It was examined by the cryo group and 
declared to be cold.  We conclude that the Q band card cage is not
connected to the right place in the F320.  

Interband collimations were determined by cycling over each band
for a few minutes in pointing mode many times.  Even K band was close 
enough to allow valid pointing determinations.  Data was taken for 
Q band but it was not working well enough to provide any solutions.  
The measured collimations relative to X band are:
Band      14            16
C     +1.85 -0.05  +0.10  0.00
K     +0.65 -0.70  +0.50 +0.30
L     -0.10 +2.30  -1.00 -0.40
These collimation corrections will be entered in the DB tomorrow.
During the collimation observations MIBs in all the downconverters
in antenna 16 stopped responding.  Rob investigated but could find
no smoking gun and reset them all.

After this Q band was tried again and it worked.  There was much 
confusion at 1630 LST caused by the switch from 3C273 to 3C345
but eventually the delays, which were off by many tens of nanoseconds
were tuned to with a few nanoseconds.  There was no time left to 
check focus or pointing.

All this was done using the table driven level setting algorithm and
no problems were seen.  Data was taken at 10 seconds and 1-2/3 seconds
and no phase spinning was seen at any band.

Next time we should continue refining Q band and probably take some
data, under carefully controlled conditions, towards understanding
all of Rick's phenomona.  Thursday Barry plans to try controlling 
VLA antennas.  We may consider using two subarrays if we are not
already confused enough.