2-page Instructions to process drift A1946 data.......
______________________________________________________

**See bottom of page for updates after late Sept run - KS 28sep04**


1) Data are assumed to be already in the form of IDL structures,
   produced by filecreator (see relevant instructions), in the
   form of .sav files such as 425206165.sav, for example, in
   a data directory such as   /share/a1946-2/data/09.06.am (at AO)
        /home/dorado5/a1946/data/09.06.am (in Ithaca)

   You are expected to be reasonably familiar with IDL.

2) Place yourself in that data directory and start IDL;
   when the IDL> prompt appears, type
        IDL> @wasinit2
   This will load the library of phil's idl routines.
   (In this and other command examples, you don;t get to type
    the 'IDL>' part).

3) Compile the modules that will calibrate, bandpass subtract and
   baseline the data. Those are in
        /share/a1946-2/idl        (in AO)
        /home/dorado5/a1946/idl (in Ithaca)
   Type
        IDL> .comp ../../idl/bpc1_rg11
        IDL> .comp ../../idl/calib5_2s
        IDL> .comp ../../idl/strip_pv2

4) Load one drift file, e.g.
        IDL> restore,'425206165.sav'
   Loading should take 20 seconds or so.

4) You can now call a module that will invoke the calibration
   procedure, the bandpass procedure, create a structure containing
   the calibration data (calib), the reduced data (dred) and other
   ancillary files. It will also automatically save the lot in a
   file named, e.g. d021131+273002.4252211577.sav
   which will stored in the same directory you're running idl on

   IDL> bpd_2s

   will do that. You will be required to provide interactive input.
   A graphic window with calON/calOF will appear, for several
   (3, for a 900sec drift) cal episodes. Use the cursor to flag the
   edges of the regions you want to include in the computation of
   total power - avoid rfi and galactic HI parts of spectrum. You'll
   do that for each beam/cal combination. Regions can be different
   from one beam/polarization to the next. Always exclude BP skirts.
   Then you'll be presented with estimates of T_cal/(calON-calOF),
   for each of the beam/pol combination. Skip (p) over board 0A
   (currently busted) and hit return on all the others, toselect
   the median value of 3 cal episodes. the interactive part will
   take 1 minute per drift; the successive processing will take
   about 6-7 minutes for the 7 beams on a good workstation.

   I have tried to consolidate a lot of functions into a single
   call, to make it easier. This has grown out of several separate
   procedures, but I tried to spare you the pain of memorizing
   too many steps and to me that of documenting (things are still
   too raw for extended documentation).
 
   Note that bpd_2s is a "cottage farm" kind of product. Many
   modifications to adapt to the condition of AO fits hdrs and
   timing problems have been cluged into this procedure in a
   ad hoc manner. It assumes that cals are fired at periods of
   275 sec, and for 2 sec, during a scan. If data do not conform
   to this mode, unexpected occurrences will take place.

5) After running bpd_2s, you can look at the processed data by
   invoking the procedure strip_pv, which will smooth the data
   and give you a view of one beam/pol at a time, by typing, e.g.

        IDL>strip_pv,dred,cont_tot,msmooth,N1=0,N2=6,GAUSAV=7,han=3,showpol=3,units=1

   This will display the strip corresponding to each beam, after
   averaging the two pols, smoothing with a han=3 in the spectral
   dimension and with a Gaussian of 7 sec HPFW in the RA dimension.
   For other options, see the comments at the beginning of the
   strip_pv2.pro file in  /share/a1946-2/idl        (in AO)
   or /home/dorado5/a1946/idl (in Ithaca)

   If you choose to use strip_pv on a formerly reduced and saved
   data file, which you are restoring for inspection, you'll have
   to re-create the cont_tot array. Do so by typing:
        IDL> cont_tot=reform(mc[1,1,*,*])

   Keep a log of the data quality of each processed drift, e.g.
   note timing problems, total power jumps (often in beams 4A
   and 5A), interesting detections, strong continuum sources,
   poor BP subtraction, rfi, etc.

-- The previous set of instructions have worked quite robustly for
   a couple of weeks. If you are not somewhat conversant with IDL,
   however, it is likely that you will stumble into problems. Data
   processing of ALFA data is certainly not yet in the "reduce by
   numbers" stage.    

****************
Modifications for late Sept run: these implement minor changes and 
corrections (by AS; ask her for details). Do:

        IDL> .comp ../../idl/bpc1_rg11
        IDL> .comp ../../idl/akbj_calib5_2s
        IDL> .comp ../../idl/akbj_strip_pv2

        IDL> .run ../../idl/akbj_bpd_2s