[evlatests] P-band performance

[Walter Brisken]

Rick asked me to look at bit at levels of P-band EVLA receivers as the 
performance on some antennas/IFs is quite bad.  There is quite a bit of 
P-band observing coming up soon.  I didn't end up making any progress due 
to other things I had to do today, but I did at least figure out where 
things are.  I hope someone can take this over from me -- I don't 
have the full expertise to deal with this issue anyway.  I will try to 
use the 30 minute dynamic slot tonight to get updated levels.  Stay 
tuned...

Currently the P-band attenuators are set just like the Cassegrain band 
attenuators -- that is A/C are automatically set and B/D use the values in 
the parameters database.  The currently active values are shown below. 
The colums labeled CC are performance measurements made by Rick (inverse 
correlation coeficients I think, referenced to a VLA antenna).  A value 
of 3 to 3.5 is normal.  Tsys scales roughly as this number squared.

Antenna	IFBOUTA	IFBRFA	CC B	IFDOUTA	IFDRFA	CC D	Comments
11	24	0	8.5	24	0	8.5
13							No P-band
14	18	10	6.0	15	7	6.5
16							No P-band
17	22	0	3.6	22	0	4.2
18	11	12	3.6	16	16	3.7
19			3.6			4.2	Attens not set
21	15	8	45	15	22	24	REALLY BAD
23	22	3	4.0	22	1	3.9
24	24	0	5.8	16	0	4.2
26	10	0	infty	18	0	infty	REALLY REALLY BAD


The A/C side has typically similar performance as the B/D side, but 
without the 4 extremely bad cases that 21,26 B/D have.  To improve on the 
A/C side needs new algorithms in the hardware.

The really bad performing antennas (26 and 21) have attenuator settings 
that don't differ much from others that have reasonable performance -- 
makes me wonder if there is something else wrong with them.

Ant 19 has no values in the parms database but performed remarkably well 
anyway.