[evla-sw-discuss] VLA antennas and the EVLA parameters database

[Ken Sowinski]

                      Old Wine into New Bottles
 		          June 28, 2006

I have started constructing a program to pour all the system parameters
from the Modcomps into the EVLA parameters database.  I want to set down
the method and rationale to give anyone who cares a chance to suggest
better approaches or techniques.

The idea is to read all the system files and generate either SQL to
insert things in the database or generate an intermediate file from
which SQL can be generated.  I will go through tha parameters grouped 
by the system file the are retrieved from.  For each required parameter
a line will be generated which contains:
ENTITY  PARAMETER  VALUE  (BAND)
The band will be included when needed, otherwise left blank.  Only
the parameters needed to run the antennas are included in this list.
The contents of the IF files will be needed later to run the correlator 
and provide T_sys calibration.  I expect this program to run whenever
we feel that the parameters database is getting stale, most likely,
whenever an array reconfiguratio occurs.  It might be worthwhile to
use the diff between the current and last output to feed the database.

Most of this is straight forward and uncontroversial.  This is not
true for the tilt parameters because of the differences in the way
azimuth and elevation are calculated in the two systems.  The important
difference is that the tilt of the pads away from the local zenith
must be explicitly accounted for in the pointing model because CALC
now provides pointing angles for the position of the antenna rather
than the center of the array.  I propose that for VLA antennas the
geometrical tilt of the pad be expressed in the pad tilts and the 
pointing model tilt from the Modcomps be properly rotated for the
arm azimuth and expressed as the antenna tilt.

SYSANTS
DCS      Expressed as a decimal number.
LOCATION the PAD converted to the newer, less informative style; e.g. W56.
KTERM    converted from nanoseconds to seconds.

SYSBASE
X, Y, Z  are rotated to Greenwhich using a longitude for the VLA of
          107.617722 degrees and converted from nanoseconds to meters.
          This is choosen because it is the longitude used by the VLA
          to calculate LST and we want the EVLA antennas to be consistent
          with the VLA. 
EWTILT   These are calculated from the B_y and B_z in the VLA coordinate
NSTILT   system.  The algorithm follows the description by Cam Wade in
          VLA Test Memo 103.  The EWtilt is the difference in longitude
          between the pad and the center of the array; similarly, for the
          NStilt and the latitude.  It is sufficient to use the small angle
          approximation and assume a single radius of curvature regardless
          of position.
          delta-Longitude = -By/(r0*cos(\phi))
          delta-Latitude  = -Bz/(r0*cos(\phi)),
          where \phi is the latitude at the center of the VLA.  Plugging
          in constants we have, when the antenna positions are expressed
          in meters:
          EWTilt = -By/1537.22
          NSTilt = -Bz/1537.22
          The signs have been verified by inspection against the pointing
          model as well as comparison with the constants Barry introduced
          for AW5 when antenna 13 was out there.

SYSPOINT
EWTILT   These are the tilts given in the antenna file rotated by the
NSTILT   azimuth of the arm the antenna resides on as defined in
          EvlaMechnicalModel.java.  The results will be inserted for
          the antenna rather than the pad.  The signs are to be those
          given for E1, E2; not the azimuth versions.
AZCENTERINGCOS    A3 and A4 are rotated like the tilts.
AZCENTERINGSIN 
PERPENDICULARITY  A5
AZENCZERO         A7
ELCENTERINGCOS    E3 and E4
ELCENTERINGSIN

SYS?ROT
AZCOLLIMATION  The sum of this number and A6 from the pointing file.
ELCOLLIMATION  The sum of this number and E5 from the pointing file
FOCUS          The focus is converted to centimeters by dividing by
                537.924 and increased by 0.55 cm for those bands at
                which the focus curve will be applied.
ROTATION       This is converted to degrees using 360/16384.