[evlatests] Level setting

[Barry Clark]

I have surveyed the effect of using the proposed T304 level
setting algorithm at the standard frequencies at all the bands.
I include the complete table at the end of this message, including
the X and C band measurements I sent along yesterday.  The overall 
conclusion is that the current software will work reasonably well 
at the high frequency bands.  There are a few wild numbers in the 
table below.  I'm not sure if I merely copied a number incorrectly 
from the screen, or if something is erratic, or if something really 
odd is going on.  Even with the odd numbers, things seem to work well 
enough to be usable.

The adjustments needed to handle the different gain between the T304 
output and the 8 bit samplers are relatively modest.  IF A in 
antenna 16 if A might profitably have 1dB less attenuation.  
IF A in antenna 14 probably needs an additional 2dB attenuation to 
get the sampler operating in a reasonable range.  The only worrisome
thing for this IF is at K band, where the RMS on the link is a bit
high, as usual, but the power presented to the T5, as represented
by the ALC voltage, is a bit low.  The suspicion is that the 50 MHz
presented to the T5 is at an exceptionally low spot in the 1 GHz 
wide spectrum driving the sampler.

L band is susceptible to interference.  One can see the total power
sloshing around by a dB or two in normal times - I didn't see anything
more than that in a few minutes of watching it.  This is little enough
that the current software would do for level setting.  However, I
suspect that rather more viscious conditions can occur even now, and 
will be the norm with the wideband frontend.  Our suggested strategy 
for dealing with it - taking the minimum detector reading in a 2 ms
sample over a few hundred of them - needs to be tested.

P band works better than I expected.  This is quite usable.

I do not understand the behavior of 4 band at all.  There is an
expected effect that the band is very sparcely occupied, and that
therefore settings that give a reasonable link RMS should give a
higher output level than something where the whole band is filled.
This should be compensated by turning down the gain in the deformatter.
But the behavior of the IF C switched power is just crazy.  The only
suggestion I can offer is that it is dominated by impulsive out-
of-band interference that splatters all over the place unless the
input amplifier is heavily padded.  

X Band, with Walter's settings from the table.
      Inat Outat inpw   outpw  RMS T5ALC T5SD
ea14-a   0    10 -23.5  -32.1  8.9 -4.3  2.80
ea14-c   0    13 -20.5  -30.8  8.8 -4.0  2.11
ea16-a   0     9 -24.9  -32.5  5.8 -1.5  1.84
ea16-c   0    15 -22.0  -34.2  5.0 -2.6  2.10
ea24-a   0    10 -22.7  -28.7 10.4 -3.9  1.7 (erratic)
ea24-c   0    10 -22.8  -33.1  5.8 -4.1  1.10
With Level setting software
ea14-a   0     7 -23.5  -29.2 11.6 -4.6  2.48
ea14-c   0    13 -20.4  -30.5  8.6 -4.0  2.22
ea16-a   0     7 -25.0  -30.8  7.2 -2.4  1.90
ea16-c   0    11 -22.0  -30.4  8.3 -3.7  2.08
ea24-a   0    10 -22.8  -28.8 10.5 -4.1  1.7 (erratic)
ea24-c   0    10 -23.0  -33.4  5.7 -4.1  2.12

C Band, with Walter's settings from the table.
ea14-a   8    15 -22.5  -30.6  9.3 -4.1  2.76
ea14-c   4    15 -17.9  -26.8 12.7 -4.3  2.69
ea16-a   4    15 -18.2  -31.7  6.4 -2.4  4.00
ea16-c   7    15 -19.6  -31.2  7.5 -3.4  4.21
With Level setting software
ea14-a   7    15 -20.7  -29.8 10.0 -4.2  2.77
ea14-c   6    17 -20.9  -30.8  7.8 -3.6  2.72
ea16-a   6    12 -20.7  -30.9  7.3 -2.6  4.03
ea16-c   7    14 -19.6  -30.4  8.0 -3.6  4.24

K Band, with Walter's settings from the table.
ea14-a   8    15 -21.7  -30.1  9.9 -4.4  2.39
ea14-c   4    15 -15.3  -26.7 13.5 -4.4  2.40
ea16-a   7    15 -19.1  -31.4  6.9 -2.7  2.05
ea16-c   9    15 -21.7  -31.3  7.3 -3.5  2.30
With Level setting software
ea14-a   7    15 -19.9  -29.3 10.8 -2.9  2.08 
ea14-c   7    15 -19.7  -29.3  9.4 -3.6  2.33
ea16-a   7    14 -19.2  -30.7  7.4 -2.9  2.07
ea16-c   8    15 -20.2  -30.4  8.3 -3.6  2.30

Q Band, with Walter's settings from the table.
ea14-a   6    15 -28.2  -30.4  9.0 -4.1  4.85
ea14-c   6    15 -22.0  -27.7 11.5 -4.2  4.78
ea16-a   6    15 -21.9  -32.4  6.0 -2.2  3.06
ea16-c   2    15 -20.7  -32.7  6.1 -3.1  3.80
With Level setting software
ea14-a   6    15 -20.1  -29.7  9.8 -4.2  4.82
ea14-c   4    19 -19.1  -29.8  8.8 -3.9  4.74
ea16-a   5    14 -20.6  -30.8  7.2 -2.7  3.08
ea16-c   2    12 -20.8  -30.1  8.5 -3.7  3.88

L Band, with Walter's settings from the table.
ea14-a   1    19 -30.2  -31.6  8.7 -4.4  4.16
ea14-c   1    15 -25.9  -27.9 12.1 -4.5  4.06
ea16-a   0     9 -30.4  -32.6  6.3 -2.8  3.45
ea16-c   0     9 -29.5  -33.0  6.1 -3.2  3.83
With Level setting software
ea14-a   4    15 -22.7  -29.7 11.1 -4.5  3.93
ea14-c   0    18 -23.6  -29.8  9.5 -4.2  3.72
ea16-a   0     7 -30.5  -30.7  7.7 -3.4  3.50
ea16-b   0     6 -29.5  -30.2  8.6 -3.9  3.76

P Band, with Walter's settings from the table. (6.25 MHz BW)
ea14-a  12     8 -41.1  -30.0  8.3 -3.6  4.05
ea14-c  10     8 -40.3  -32.5  5.9 -4.5  5.41
With Level setting software
ea14-a   0    19 -25.0  -29.4  8.7 -4.5  4.04
ea14-c   0    14 -27.2  -29.1  9.7 -4.8  5.41

4 Band, with Walter's settings from the table. (0.78 MHz BW)
ea14-a  12     8 -41.1  -34.0  6.9 -5.1  0.45
ea14-c  10     8 -40.3  -36.7  4.3 -5.0  4.90
With Level setting software
ea14-a   0    16 -25.1  -30.6 11.0 -5.1  0.58
ea14-c   0    10 -27.2  -29.8 10.4 -5.1  0.40