[evlatests] Birdies in 14a

[Mike Revnell]

I went out this morning to look into the problem.

The deformatter in 14a was borrowed from the antenna 13 set to replace one 
that was having optical connector problems. There is a jumper at the line 
driver output that goes to the vla baseband system that wasn't installed. 
I installed the jumper and the birdie went down by over 30dB.

Some background on this jumper:

The line to the VLA baseband system is driven directly by the high speed 
DAC output. This output is a differential current output and is coupled to 
the line by a transformer. It is difficult to know beforehand how the 
output side of the transformer should be treated so I put a component 
there which could be left open, shorted or connected with a resistor or 
capacitor as determined in tests. The boards are delivered from the board 
manufacturer without the parts installed.

In our system they seem to want to be installed. When assembling the 
boards I don't always remember to install them.

I installed the jumpers in the rest of the ant 13 board set.

Here are the results of measurements I made at the ant 14 T5 test points.

Observed with the Agilent spectrum analyzer in the central elec. room.
At 100 Hz RBW the broadband noise level was -70dBm.

BEFORE INSTALLING JUMPER
IF  100HzRBW                     10HzRBW
A   707KHz -50dBm 695KHz -50dBm  706 KHz -40dBm 694KHz -40dBm
B   Not seen                     695 KHz -70dBm
C   687KHz -60dBm 697KHz -70dBm  Didn't look
D   Not seen                     Not seen

After installing the jumper in IF A the lines went down to about -73 dBm 
at 10Hz RBW.

The decrease in power at narrower bandwidth indicated to me the lines are 
beginning to be resolved. The frequency and frequency variation are 
consistent with the switching regulators used to reduce 48v to the 
internal operating voltages.

Mike Revnell
National Radio Astronomy Observatory
Socorro, New Mexico
mrevnell at nrao.edu
revnell at ieee.org