Automatic Inclinometer Adjustment of the Pointing Model	    15 Feb 02

    The software is now in place to make inclinometer measurements during 
slews and update the pointing model with the results.  This is done via
two programs, RT_GET_INC and FIT_INC_REALTIME.  RT_GET_INC runs on Kronen
and collects inclinometer data when the antenna velocity exceeds 50% of 
slew speed.  The data is recorded on disk if the total distance traveled is
greater than 100 degrees.  When the data is recorded, RT_GET_INC sends a 
message to FIT_INC_REALTIME.  This program runs as a batch job on Frankenheim.
It fits the data recorded and calculates the West and North components of 
the azimuth axis tilts, IncW and IncN.  It uses these to make a running 
average of IncW and IncN which are recorded on disk.  Program Control reads
the results and stores them in global memory for program Fahren to use in
the pointing model.

    The accuracy of the fits is better the larger the slew.  A weight is 
assigned to the fit based upon the slew distance.  The weight is used in the
running average calculation to prevent low accuracy fits from spoiling the
average.  The weight for slews greater than 160 degrees is high.  Below 140
degrees it begins decreasing and is practically zero below 100 degrees.

    If Control notices that it has been 24 hours since the last slew of 160
degrees or greater, it begins nagging for a slew with a message in the 
error log: "SLEW >160 OLD" (once every 30 minutes).  If it has been 48 hours
since the last slew, the messages changes to "NEED SLEW >160".

    A SLEW command has been added to OBST to make it easy for the observer 
to request a long slew.  Its syntax is

	    SLEW [<distance>] [H or Z]

where <distance> is the slew length in degrees (negative means CCW), and the
H or Z specifies the elevation: zenith (85) or horizon (3).  If no <distance>
is specified, + or - 175 degrees is chosen, with the sign being the one that
leaves the antenna closest to Azimuth=180.  If neither H or Z is specified, Z
is assumed. (One is allowed to omit the <distance> and yet specify the H or Z.)
If the specified <distance> would exceed the wrap limit of the telescope, the
opposite sign will be chosen.  If neither sign is possible, the largest
possible <distance> is chosen.  If the elevation chosen is within 35 degrees of
the sun's elevation, the sun's elevation, + or - 35 degrees is chosen.

    If Control finds that IncW or IncN has changed by more than one arcsec from
the previous value, it posts a warning message to the error log: "AUTO TILT 
JUMP".  

    The current values of IncW and IncN parameters are given at the end of 
the line on the SDH display that contains the focus positions.  The OBSINP
command, TILT, turns on or off the automatic update (when the next observation
is begun).  To check the status in OBST, type

		OBSINP
		SHOW DP1

The TILT mode (AUTO or FILE) will be displayed followed by the current tilt 
values.  Type TILT FILE to use the tilts from the POINTCON.DAT file and 
TILT AUTO to use the running average from the inclinometer.  (If you change
the mode, it doesn't take effect until the next observation.  And the values 
on the OBSINP display won't change until you do another SHOW DP1.)

Result files:

The IncW and IncN values read by Control are stored in 

    		kdata:[smt.inc]cur_tilt.dat

and the history of realtime inclinometer fits are stored in

    		kdata:[smt.inc]rtinc.sum

The former stores 

<day last> <day big>  IncN  IncW <last file>

Where <day last> is the day of the last fit (in days since Sept 1), <day big>
is the day of the last fit to a slew bigger than 160 degrees, IncN and IncW
are the running average north and west components of the azimuth axis tilt
and <last file> is the name of the file that contains the inclinometer 
measurements.  (These files are first stored in directory kdata:[smt.inc]
by RT_GET_INC and are moved to kdata:[smt.inc.save] after FIT_INC_REALTIME
fits their data.)

    The inclinometer IncN and IncW are not the same as used for pointing.  They
are systematically different from these, presumably due to local gravitational
effects of the mass of Mt Graham and a figure of the earth that deviates from
the standard flattened spheroid.  (The latter would be due to the distribution
of mass of the Southwest US.)  To convert from inclinometer tilts to pointing
tilts, the following (empirical) offsets are used: 

		    	Spring 2001 (?) 
		    IncW(radio) = IncW(Incl) + 1.5"
		    IncN(radio) = IncN(Incl) + 6.5"


To see the header line and the last 10 entries of the rtinc.sum file, type 
rt_inc in a Jever smtobs window or a Kronen smtsys window.  (To prevent line
wrapping, do this in a window with >=143 columns.)  The most relevant columns
are the Date (days since Sept 1), Deg. (slew length), IncN and IncW (from the
last fit), and Mean IncN & W (the running average).