[evlatests] Ant 14 & 16 Hardware

[Rob Long]

We had an interesting day investigating EVLA hardware.  The plan was to 
radiate a CW signal to both antennas and look at very close (sub 1 Hz) 
offsets from the carrier and see what might be happening to the signal. 
  All previous tests were done from 10 Hz offset and greater with 1 Hz 
RBW  with no extreme signal degradation or excessive phase noise measured.

Today, we radiated at 4.898 GHz, used C-band receivers, and standard LO 
frequencies for IF-A (L301 & L302).  After driving ant 16 into its lower 
El limit, Ken informed us what we did wrong while using ETS.  Once both 
antennas were pointed at the building, we enabled the output of the 
signal generator and immediately saw the CW in the output of the T5-As 
for both antennas.

After playing with levels to optimize S/N (and properly drive our 
mixer), we combined the signal from both T5-As into the mixer and sent 
the output of the mixer into the dynamic signal analyzer.  We were set 
up to monitor frequency from DC to 1 Hz and to monitor peak-peak voltage 
output of the mixer.

We saw a spur at 0.8 Hz (-36 dBm) that appeared to be slowly drifting. 
While looking at the scope, we decided to narrow the bandwidth of the T4 
filter to help with extraneous noise and used the 6.25 MHz setting. 
While doing this, antenna 16 decided to time out and autostow.  Once we 
realized what was going on, we repointed ant 16 and hit e-stops.  When 
we reacquired the signal, the spur was now at 0.25 Hz and drifting 
higher in freq.  At this point, we ran over to the cafeteria for a quick 
lunch.

When we returned from lunch, the spur was sitting at 0.77 Hz and the 
pk-pk voltage was over 400 mV.  We wondered if the fiber could be the 
culprit and messed with the central rack fiber and connectors.  The spur 
did not change (other than the slow drift) but pk-pk voltage changed to 
less than 50 mV.  Jim tightened and loosened connectors, wiggled fibers, 
etc. and several jumps back and forth in pk-pk voltage occurred.  It 
finally settled in at just under 200 mV and remained fairly constant.

At this point, we decided to check and disable phase switching and the 
spur went away.  For a sanity check, I commanded the L302 in ant 16 off 
by 1 Hz.  We immediately went back to the correct frequency and the beat 
note went to 2 Hz!!  Thoroughly confused, I commanded the L302 off by 
100 MHz and back to see if it was not really locked and needed to be 
reset.  We now saw a sub 0.1 Hz beat note and a corresponding slow sign 
wave in the pk-pk voltage.

Checking with Ken to disable the fringe rotator made no difference and 
the beat note started drifting higher in freq. very slowly (many seconds 
period).  We were obviously not at exactly the same frequencies in both 
antennas and things were pointing at the L302s.  At this time, we turned 
the antennas over to Ken and Rick and planned to make some identical 
measurements with the L302s in the lab.

As a side note, we were informed that L302s in both ant 14 & 16 had to 
have their MIBs rebooted before the antennas would fringe.

Rob Long