[evlatests] mysteriouser and mysteriouser

[Ken Sowinski]

                      November 15 2005, EVLA Tests

An hour and a half was available this morning for PT checkout.
Pietown seems to be in good shape so I prevailed upon Mark to
use the time to track down the problem discovered last Friday.
I tried to find where the fringes to VLA antennas had gone.  
It seems that the main culprit was the deformatters, but there 
may be more things going on than I have yet discovered.

Lessons:
1.  The required command to select High or Low band seems to have
changed state.  A '1' is required now; last week it was a '0'.

2.  Even in the state that produces correlated signal the results
do not yet look normal.  The main problem is that the correlation
coefficient is too small.  This may be because the deformatter/
transition module is doing something wrong, or because there are
other problems in the system.  My impression, at all bands, is 
that 16 was worse behaved than 14 be a factor of 5-10 in power.

3.  Delays ssem to have changed a lot since last week.  Even stranger
the delay errors measured at 6 MHz were not at all the same as those
measured at 50 MHz.  Narrowing the BW for the two antennas seen on
D10 did not make much difference to the amplitude, so there must be
something happening that is not a simple delay error.

4.  In the middle of all this antenna 16's subreflector wandered
off to where it ought not to be.

5. 16B was seen to have unstable amplitudes.  The variations were 
much larger when looking at 8 x 16 than when looking at 14 x 16.
There were about four states covering about a factor of 50 in power.
The ALC, TP and SD seen at the backend were all stable while this
was happening.


Tomorrow, unless I someone suggests a better plan, I will reboot
one of the deformatters to see what happens.  Any other suggestions
are welcome.