[evlatests] More results, from more tests ...

[Rick Perley]

     I've analyzed the L-band data from yesterday afternoon.  This test 
did a simple 'back and forth', within the main beam, at L-band.  The 
phase/delay stepsize was 10 seconds, and the antennas moved at '4 x 
Sidereal (1 arcminute/second) ' for the RA cuts, and 2 x Sidereal (30 
arcseconds/second) for the Dec cuts.

     The data were calibrated as described earlier.  Referenced pointing 
was claimed to have been done, but for reasons explained below, there is 
reason to believe it was not applied properly.

     I give the beam-crossing deviations (in seconds of time) below.

Ant  W to E   E to W   N to S   S to N
---------------------------------------------------
1        0.16      -0.21      -0.53      0.52
2        0.14      -0.12      -0.03      0.00
3        0.01       0.03       0.12      -0.11
5       -0.12       0.06       0.18     -0.18
6        0.14      -0.13       0.00      0.06
7       -0.05       0.04       0.16      0.04
8        0.25      -0.34       0.28     -0.20
9       -0.07       0.03       0.14     -0.18
10      0.03      -0.08       0.17      0.07
11     -0.02      -0.04       0.10     -0.22
12     -0.07       0.13       0.10     -0.05
13      0.07      -0.02      -0.04     -0.07
14     -0.09       0.26       0.08     -0.09
15     -0.16       0.17      -0.24      0.20
16     -0.02      -0.06       0.14     -0.14
17      0.05        0.02     -0.10     -0.08
18      0.00      -0.04       0.14     -0.09
19      0.06      -0.10       0.51     -0.58
20      0.10       0.00      -0.19      0.02
22      0.10      -0.06       0.06      0.10
23     -0.14       0.15      -0.11      0.30
24     -0.12       0.06       0.06     -0.20
25     -0.09       0.11       0.24     -0.24
26     -0.06       0.15      -0.11      0.11
27      0.01      -0.20      -0.16      0.03
28        xxx         xxx      -0.94     0.97
------------------------------------------------------

     It will quickly be noted that virtually all offsets larger than 0.2 
seconds of time have opposite signs for the reversed motions. For 
example, ea28 is 'early' by one second on the North to South scan, and 
is 'late' by one second on South to North.

     This a clear sign of a pointing error.  For ea28, this is easy to 
understand -- the antenna joined the array late, and missed the 
referenced point scan -- which was at the beginning.  One second of time 
is 30 arcseconds -- a perfectly normal offset for late afternoon.

     But if this is so, and if the remaining antennas did get their 
referenced pointing done, and applied, why are so many of these 
time-crossing so large?  They should all be less than 0.2 seconds of 
time (equal to 12 arcseconds for the RA motions, half that for Dec 
motions).

     Conclusion:  If the larger time offsets are indeed due to 
uncorrected pointing residuals, the residual time crossings are 
typically 0.1 seconds of time -- a few arcseconds.

     This test needs to be repeated at X-band, where we have more 
confidence in the referenced pointing, and the beam shape is much narrower.