Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.naic.edu/~astro/spectral_line/handbook/app/node10.html
Дата изменения: Wed Mar 20 18:34:05 2002
Дата индексирования: Sun Dec 23 03:51:03 2007
Кодировка:
Поисковые слова: http astrokuban.info astrokuban

OLD ANALYZ instruction sequences HI in galaxies

For use with data taken BEFORE January 1, 2000

This appendix contains example analyz sequences used to look at HI spectra.

This routine takes the average of the (gain/za corrected) dumps in each scan, and analysis is done on the individual correlator sections. Thus if you have the correlators overlapping in frequency, and the galaxy therefore shows up in more than one correlator this routine will only look at one `detection' at a time. The commands to by typed are in boldface, followed by a description of the command. For help with these routines contact Karen O'Neil (koneil@naic.edu). All this assume you have run once analyz before. If you have not, please go to http://www.naic.edu/$\sim$astro/spectral_line/analyz/old and load HI_before01jan00.cmd, mjsr.o, and baseh.o into you analyz directories. Additionally, if you have not run analyz before, instruction (2) should read:

(3)	analyz
		$<proj\:name>$ $<num\:chan>$ 800	Type your project name,
			"num chan" = total no. of channels needed.
		corv1	This chooses the correct correlator libraries
		$\sim$HI_before01jan00.cmd	This puts the libraries you need into your analyz directory
		autoexecute
(1)	cd analyz		change to your ANALYZ directory
(2)	analyz		starts the analyz program
		$<proj\:name>$	Type your project name
		corv1	This insures the correct correlator libraries are attached
(3)	attach $<data>$ data		attaches the desired data file
(4)	corget $<scan\:num>$		a temporary fix...
(5)	setfrq $<fr1>\:<fr2>\:<fr3>\:<fr4>$		set sbc freqs. to fr1, etc.
(6)	redln $<rec>$ $<n>$		scans n records, starting with rec, corrects for
	(redlw $<rec>$ $<n>$)		gain/za variations, ultimately stores the data in reg. 43
(7)	s $<s1>$ $<s2>$;hor;plot		selects the region of interest (i.e. where a galaxy
			images lies), plots the data
(8)	getres $<low>$ $<high>$		gets vel widths, etc.
			$low$ & $high$ are the channel numbers which
			encompass the galaxy

DESCRIPTION of the above commands:

• (1) cd analyz: By changing into your "analyz" directory, you are insuring that analyz will be able to find your libraries, any necessary fortran files, etc.

• (2) analyz...: Entering the correct project name here ($<proj\:name>$) insures that analyz recognizes the correct number of store/recall registers and that analyz finds the libraries associated with your work. These files are located in your analyz directory, and are labeled, i.e. a9999anz.anz and a9999lib.anz, respectively.

• (3) attach $<data>$ data: This attaches the (old) data to analyz. After you "mvdata" while observing, your data is typically placed in a files called "/share/olcor/corfile.$<data$>.$<proj>$.n", i.e. if the data for project a9999 was taken on March 05, 2000, the file would be "/share/olcor/corfile.05mar00.a9999.1.

• (4) corget $<scan\:num>$: You need to repeat this step, and step #6, only whenever the frequencies, bandwidths, etc. at which your data was taken change, and when you first start analyz.

• (5) setfrq $<fr1>\:<fr2>\:<fr3>\:<fr4>$: Again, you only need to repeat this step whenever he frequencies, bandwidths, etc. at which your data was taken change, and when you first start analyz. This step determines the correct frequencies, bandwidths, etc. for you data and stores the result in register 8. (NOTE: this only works if your A & B polarizations have the same frequency settings.)

• (6) redln $<rec>$ $<n>$: This first stores the individual ON/OFF - 1 data dumps into registers 100 - 100+$<num\:rec>$ - 1. It then does a rough baseline correction, corrects each ON/OFF - 1 for gain/za variations and converts temperature to Janskies (using the standard SEFD curves), and stores those in registers 100+$<num\:rec>$ - 100+2*$<num\:rec>$ - 1. Next, the average of all the (ON/OFF-1) records is found, and stored in register 100+2*$<num\:rec>$. Finally, the A & B polarizations are averaged, and those results are stored both in register 100+2*$<num\:rec>$+1, and in register 43 (where the next set of routines expects to find the data. (NOTE: Make your you have 100+2*$<num\:rec>$+1 registers available!)

• (7) s....: At this point, you need to figure out what range of channel numbers you wish to use to analyz your data. You need a range big enough so that you can obtain an accurate baseline, yet small enough that you can readily determine the sides of your galaxy.

• (8) getres $<low>\:<high>$: $<low>$ & $<high>$ are the channel number range chosen in the last step. This routine does the following:
  • setvel: Uses the "optical" definition to convert frequencies to velocity, and then stores the velocities in register 41. The data is then "sectioned" for the rest of the analysis.
  • baseh -5: This runs the baseline reduction routine, allowing you to choose which order fit you would like (from 1 to 5). The baseline is then subtracted from the data, and the results are stored in register 47.
  • msrtest: This runs the msrj program, and outputs the following results:
    Line 1 - The area, central velocity, and width at 50% of the mean flux.
    Line 2 - The area, central velocity, and width at 20% of the mean flux.
    Line 3 - The area, central velocity, and width at 50% of the peak flux.
    Line 4 - The area, central velocity, and width at 20% of the peak flux.
    Line 5 - The area, central velocity, and width at 20% of each of the two horns,
    searched from the inside out.
    NOTE: Ignore the minus signs on the widths and areas - they come about because the data is flipped from what the msrj program expects (so the high velocity is subtracted from the low, instead of vice-versa.)

------------------

EXAMPLE using OLD routines (this data will be available for public use):

FIRST time:

        % cd analyz
        % analyz
        Enter your initials  [array size and number of STO/RCL registers optional]
        > temp 16284 800
        File ANZ_BASE/site/syslib.anz attached as SYSLIB.

        Enter <CR> or select an FLIB from the following:
         corv1    CORLIB for data prior to:07mar00
         corv2    CORLIB for data after   :06mar00
        FLIB> corv1
        File ANZ_BASE/site/corv1lib.anz attached as FLIB.

        >>> Creating a file for the stack and sto/rcl registers <<<
        Attaching temppanz.anz as STO/RCL file.
        Your arrays are 16384 elements long, and you have 800 sto/rcl registers.
        None of your sto/rcl registers are kept in main memory.
        Loading... PL X Y Z T HDR 
        File templib.anz attached as ULIB1
 
        Welcome to Analyz, version 5.16 patchlevel 2
 

        >~hi_before01jan00.cmd
        Error 144: ANALYZ fault: unrecognized command structure

        >exit

        STOP: ANALYZ exit
         Note: Following IEEE floating-point traps enabled; see ieee_handler(3M): 
         Overflow;  Division by Zero;  Invalid Operand; 
         Sun's implementation of IEEE arithmetic is discussed in 
         the Numerical Computation Guide.

        %analyz

        Enter your initials  [array size and number of STO/RCL registers optional]
        >temp

        Enter <CR> or select an FLIB from the following:
         corv1    CORLIB for data prior to:07mar00
         corv2    CORLIB for data after   :06mar00
        FLIB>corv1

        File ANZ_BASE/site/corv1lib.anz attached as FLIB.
 
        Attaching tempanz.anz as STO/RCL file.
        Your arrays are 16384 elements long, and you have 800 sto/rcl registers.
        None of your sto/rcl registers are kept in main memory.
        Loading... PL X Y Z T HDR 
        File templib.anz attached as ULIB1.
 
        Welcome to Analyz, version 5.16 patchlevel 2
 

        File /usr/local/phil/philib.anz attached as ULIB2.
        >

        > attach "/proj/a1366/corfile.18feb00.a1366.2" data
        File /proj/a1366/corfile.18feb00.a1366.2 attached as DATA.
        Status = OLD; access = SEQUENTIAL; form = UNFORMATTED
 
        > corget
        loading via: ANZ_BASE/tools/ld -N -x -A /usr/local/bin/analyz.exe -T 0359c00 -o a.out
        ANZ_BASE/external/extern_main.o  ANZ_BASE/external/libextern.a -LANZ_BASE/external -l
        F77 -lm -lc
        New program text loaded successfully.
        Created new array named @ci
        Created new array named @frqa
        Created new array named @frqb
           Scan    grp   time  bw integ  bd lag   pwrRatio         power        attn
                   num  ofRec Num  secs  id cfg    ratio          counters       db
          4900128     1  55050  1   1.000 6   8  1.067  0.995    56629    49613 -1 -1  1.0
          4900128     1  55050  1   1.000 7   8  0.957  1.038    42919    48003 -1 -1  1.0
          4900128     1  55050  1   1.000 8   8  1.088  0.968    52573    49050 -1 -1  1.0
          4900128     1  55050  1   1.000 9   8  0.929  1.100    44908    53388 -1 -1  1.0

        >setfrq 1262.5 1307.5 1352.5 1397.5

        >redlw 132 120

                .....

            mean     |   N   |     rms     |   skewness  |  kurtosis
          4.9025E-02 | 16384 |   2.215     |   107.4     |  1.2964E+04
        on/off -1 raw in 100 -   219.0
           Scan    grp   time  bw integ  bd lag   pwrRatio         power        attn
                   num  ofRec Num  secs  id cfg    ratio          counters       db
          4900132     1  60710  1   1.000 6   8  0.902  1.061    49512    52429 -1 -1  1.0
          4900132     1  60710  1   1.000 7   8  1.017  1.111    45853    51781 -1 -1  1.0
          4900132     1  60710  1   1.000 8   8  0.926  1.035    46470    53400 -1 -1  1.0
          4900132     1  60710  1   1.000 9   8  0.983  0.930    47183    46043 -1 -1  1.0
        Created new array named @tmp2
        Created new array named @tmp4
        on/off -1 corr in   220.0     -   339.0
        on/off -1 corrected average in   340.0
        on/off -1 corr, AB avg'd in   341.0        42.00
            mean     |   N   |     rms     |   skewness  |  kurtosis
         -5.9280E-05 |  8192 |   2.108     |   4.355     |   1521.
        rms subtracted data in 43
 
        > rcl 43;ver 0 0.05;plot

        > hor 7000 7500;plot

        > getres 7000 7500

        loading via: ANZ_BASE/tools/ld -N -x -A /usr/local/bin/analyz.exe -T 0579400 -o a.out baseh.o^M
          -LANZ_BASE/external -lF77 -lm -lc
        New program text loaded successfully.

         Set base regions, with crosshair:
         Hit space key after each positioning,
         DO NOT touch mouse button!
         Hit Q after last (rightmost) limit.
        Limits are:   2  481  92  293  0  0  0  0  0  
        Are limits OK? (y/n) [y] y
 
         Order     rms    Ftest
           1      0.002     0.367
           2      0.002     3.652
           3      0.002     0.611
           4      0.002     0.173
           5      0.002     0.197
        What order do you want?  1
        loading via: ANZ_BASE/tools/ld -N -x -A /usr/local/bin/analyz.exe -T 059e000 -o a.out msrj.o^M
          -LANZ_BASE/external -lF77 -lm -lc
        New program text loaded successfully.
        ow flag LEFT side chan for area bounds.
         Hit any key.
         Flag RIGHT side channel.
         Hit any key.
        Happy with the choice of bounds?  (y/n) [y] y
         area(150, 241) =-.6971E+01 Center = 4751.21834    Width(161, 238) = -411.97125
         area(150, 241) =-.6971E+01 Center = 4733.19330    Width(154, 241) = -527.34014
         area(150, 241) =-.6971E+01 Center = 4747.34298    Width(169, 232) = -340.31012
         area(150, 241) =-.6971E+01 Center = 4754.72229    Width(156, 241) = -457.31378
         area(150, 241) =-.7007E+01 Center = 4763.60470    Width(152, 243) = -481.15713
         area = -.6971E+01 Ok? (y/n or e (for exit))e
        MSR Done...
        >

________________________________________________

If RFI is a problem with your data, one way of reducing its effects is to take have a large number of dumps and take the median, rather than the average of the dumps (i.e. set the dumplength to 6s and still take 5 minute scans). In this case you can replace "(7) doit $<$rec$>$ $<$num$>$" with "(7) mednf $<$reg num$>$ $<$rec$>$ $<$num$>$", where register number should be above 100. This will dump the median (rather than the average) of the $<$num$>$ scans into the correct registers. Another alternative, which only works for the l-wide receiver, is to run "(7) medlwf $<$reg num$>$ $<$rec$>$ $<$num$>$", which will output the data in Jy into the appropriate registers. In this case the results given after the "getfcts" command are already in Jy and do not need to be converted.